Prognostic models of diabetic microvascular complications: a systematic review and meta-analysis

Diabetes-Related Macrovascular Complications Are Associated With an Increased Risk of Diabetic Microvascular Complications: A Prospective Study of 1518 Patients With Type 1 Diabetes and 20 802 Patients With Type 2 Diabetes in the UK Biobank

BACKGROUND

Diabetic vascular complications share common pathophysiological mechanisms, but the relationship between diabetes-related macrovascular complications (MacroVCs) and incident diabetic microvascular complications remains unclear. We aimed to investigate the impact of MacroVCs on the risk of microvascular complications.

METHODS AND RESULTS

There were 1518 participants with type 1 diabetes (T1D) and 20 802 participants with type 2 diabetes from the UK Biobank included in this longitudinal cohort study. MacroVCs were defined by the presence of macrovascular diseases diagnosed after diabetes at recruitment, including coronary heart disease, peripheral artery disease, stroke, and ≥2 MacroVCs. The primary outcome was incident microvascular complications, a composite of diabetic retinopathy, diabetic kidney disease, and diabetic neuropathy. During a median (interquartile range) follow-up of 11.61 (5.84-13.12) years and 12.2 (9.50-13.18) years, 596 (39.3%) and 4113 (19.8%) participants developed a primary outcome in T1D and type 2 diabetes, respectively. After full adjustment for conventional risk factors, Cox regression models showed significant associations between individual as well as cumulative MacroVCs and the primary outcome, except for coronary heart disease in T1D (T1D: diabetes coronary heart disease: 1.25 [0.98-1.60]; diabetes peripheral artery disease: 3.00 [1.86-4.84]; diabetes stroke: 1.71 [1.08-2.72]; ≥2: 2.57 [1.66-3.99]; type 2 diabetes: diabetes coronary heart disease: 1.59 [1.38-1.82]; diabetes peripheral artery disease: 1.60 [1.01-2.54]; diabetes stroke: 1.50 [1.13-1.99]; ≥2: 2.66 [1.92-3.68]). Subgroup analysis showed that strict glycemic (glycated hemoglobin <6.5%) and blood pressure (<140/90 mm Hg) control attenuated the association.

CONCLUSIONS

Individual and cumulative MacroVCs confer significant risk of incident microvascular complications in patients with T1D and type 2 diabetes. Our results may facilitate cost-effective high-risk population identification and development of precise prevention strategies.

A multimodal approach using fundus images and text meta-data in a machine learning classifier with embeddings to predict years with self-reported diabetes - An exploratory analysis

AIMS

Machine learning models can use image and text data to predict the number of years since diabetes diagnosis; such model can be applied to new patients to predict, approximately, how long the new patient may have lived with diabetes unknowingly. We aimed to develop a model to predict self-reported diabetes duration.

METHODS

We used the Brazilian Multilabel Ophthalmological Dataset. Unit of analysis was the fundus image and its meta-data, regardless of the patient. We included people 40 + years and fundus images without diabetic retinopathy. Fundus images and meta-data (sex, age, comorbidities and taking insulin) were passed to the MedCLIP model to extract the embedding representation. The embedding representation was passed to an Extra Tree Classifier to predict: 0-4, 5-9, 10-14 and 15 + years with self-reported diabetes.

RESULTS

There were 988 images from 563 people (mean age = 67 years; 64 % were women). Overall, the F1 score was 57 %. The group 15 + years of self-reported diabetes had the highest precision (64 %) and F1 score (63 %), while the highest recall (69 %) was observed in the group 0-4 years. The proportion of correctly classified observations was 55 % for the group 0-4 years, 51 % for 5-9 years, 58 % for 10-14 years, and 64 % for 15 + years with self-reported diabetes.

CONCLUSIONS

The machine learning model had acceptable accuracy and F1 score, and correctly classified more than half of the patients according to diabetes duration. Using large foundational models to extract image and text embeddings seems a feasible and efficient approach to predict years living with self-reported diabetes.

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.

The Reporting Quality of Machine Learning Studies on Pediatric Diabetes Mellitus: Systematic Review

BACKGROUND

Diabetes mellitus (DM) is a major health concern among children with the widespread adoption of advanced technologies. However, concerns are growing about the transparency, replicability, biasedness, and overall validity of artificial intelligence studies in medicine.

OBJECTIVE

We aimed to systematically review the reporting quality of machine learning (ML) studies of pediatric DM using the Minimum Information About Clinical Artificial Intelligence Modelling (MI-CLAIM) checklist, a general reporting guideline for medical artificial intelligence studies.

METHODS

We searched the PubMed and Web of Science databases from 2016 to 2020. Studies were included if the use of ML was reported in children with DM aged 2 to 18 years, including studies on complications, screening studies, and in silico samples. In studies following the ML workflow of training, validation, and testing of results, reporting quality was assessed via MI-CLAIM by consensus judgments of independent reviewer pairs. Positive answers to the 17 binary items regarding sufficient reporting were qualitatively summarized and counted as a proxy measure of reporting quality. The synthesis of results included testing the association of reporting quality with publication and data type, participants (human or in silico), research goals, level of code sharing, and the scientific field of publication (medical or engineering), as well as with expert judgments of clinical impact and reproducibility.

RESULTS

After screening 1043 records, 28 studies were included. The sample size of the training cohort ranged from 5 to 561. Six studies featured only in silico patients. The reporting quality was low, with great variation among the 21 studies assessed using MI-CLAIM. The number of items with sufficient reporting ranged from 4 to 12 (mean 7.43, SD 2.62). The items on research questions and data characterization were reported adequately most often, whereas items on patient characteristics and model examination were reported adequately least often. The representativeness of the training and test cohorts to real-world settings and the adequacy of model performance evaluation were the most difficult to judge. Reporting quality improved over time (r=0.50; P=.02); it was higher than average in prognostic biomarker and risk factor studies (P=.04) and lower in noninvasive hypoglycemia detection studies (P=.006), higher in studies published in medical versus engineering journals (P=.004), and higher in studies sharing any code of the ML pipeline versus not sharing (P=.003). The association between expert judgments and MI-CLAIM ratings was not significant.

CONCLUSIONS

The reporting quality of ML studies in the pediatric population with DM was generally low. Important details for clinicians, such as patient characteristics; comparison with the state-of-the-art solution; and model examination for valid, unbiased, and robust results, were often the weak points of reporting. To assess their clinical utility, the reporting standards of ML studies must evolve, and algorithms for this challenging population must become more transparent and replicable.

Systematic reviews of machine learning in healthcare: a literature review

INTRODUCTION

The increasing availability of data and computing power has made machine learning (ML) a viable approach to faster, more efficient healthcare delivery.

METHODS

A systematic literature review (SLR) of published SLRs evaluating ML applications in healthcare settings published between1 January 2010 and 27 March 2023 was conducted.

RESULTS

In total 220 SLRs covering 10,462 ML algorithms were reviewed. The main application of AI in medicine related to the clinical prediction and disease prognosis in oncology and neurology with the use of imaging data. Accuracy, specificity, and sensitivity were provided in 56%, 28%, and 25% SLRs respectively. Internal and external validation was reported in 53% and less than 1% of the cases respectively. The most common modeling approach was neural networks (2,454 ML algorithms), followed by support vector machine and random forest/decision trees (1,578 and 1,522 ML algorithms, respectively).

EXPERT OPINION

The review indicated considerable reporting gaps in terms of the ML's performance, both internal and external validation. Greater accessibility to healthcare data for developers can ensure the faster adoption of ML algorithms into clinical practice.

Plasma thrombin-activatable fibrinolysis inhibitor and the 1040C/T polymorphism are risk factors for diabetic kidney disease in Chinese patients with type 2 diabetes

Background

Inflammatory and hemostatic disorders in diabetic microangiopathy (DMA) can be linked to thrombin-activatable fibrinolysis inhibitor (TAFI) and its own gene polymorphisms. Thus, the study aimed to investigate the associations of plasma TAFI and gene polymorphisms with DMA in Chinese patients with type 2 diabetes (T2D).

Methods

Plasma TAFI of 223 patients with T2D was measured, and the genotypes and alleles of the 1040C/T, 438G/A, and 505G/A polymorphisms of the TAFI gene were analyzed. A ROC curve was constructed to evaluate the identifying power of TAFI between patients with T2D and DMA, and logistic regression analysis was used to observe the correlation of plasma TAFI and gene polymorphisms with the risk for DMA.

Results

Plasma TAFI was higher in patients with DMA than in patients with only T2D (p < 0.05). TAFI exhibited the largest area under ROC in identifying diabetic kidney disease (DKD) from only T2D (0.763, 95% CI [0.674-0.853], p < 0.01), and adjusted multivariate analysis showed a high odds ratio (OR: 15.72, 95% CI [4.573-53.987], p < 0.001) for DKD. Higher frequencies of the CT genotype and T allele of the 1040C/T polymorphism were found in DKD compared with only T2D (respectively p < 0.05), and the CT genotype exhibited a high OR (1.623, 95% CI [1.173-2.710], p < 0.05) for DKD. DKD patients with the CT genotype had higher plasma TAFI levels, while T2D and DKD patients with CC/TT genotypes had lower plasma TAFI levels.

Conclusion

Plasma TAFI and the CT genotype and T allele of the 1040C/T polymorphism are independent risk factors for DKD in Chinese T2D patients.

DNA methylation age acceleration is associated with risk of diabetes complications

BACKGROUND

Patients with Type 2 diabetes mellitus (T2D) are at risk for micro- and macrovascular complications. Implementable risk scores are needed to improve targeted prevention for patients that are particularly susceptible to complications. The epigenetic clock estimates an individual's biological age using DNA methylation profiles.

METHODS

In this study, we examined older adults of the Berlin Aging Study II that were reexamined on average 7.4 years after baseline assessment as part of the GendAge study. DNA methylation age (DNAmA) and its deviation from chronological age DNAmA acceleration (DNAmAA) were calculated with the 7-CpG clock (available at both timepoints, n = 1,071), Horvath's clock, Hannum's clock, PhenoAge and GrimAge (available at follow-up only, n = 1,067). T2D associated complications were assessed with the Diabetes Complications Severity Index (DCSI).

RESULTS

We report on a statistically significant association between oral glucose tolerance test results and Hannum and PhenoAge DNAmAA. PhenoAge was also associated with fasting glucose. In contrast, we found no cross-sectional association after covariate adjustment between DNAmAA and a diagnosis of T2D. However, longitudinal analyses showed that every additional year of 7-CpG DNAmAA at baseline increased the odds for developing one or more additional complications or worsening of an already existing complication during the follow-up period by 11% in male participants with T2D. This association persisted after covariate adjustment (OR = 1.11, p = 0.045, n = 56).

CONCLUSION

Although our results remain to be independently validated, this study shows promising evidence of utility of the 7-CpG clock in identifying patients with diabetes who are at high risk for developing complications.

Environmental exposures in machine learning and data mining approaches to diabetes etiology: A scoping review

BACKGROUND

Environmental exposures are implicated in diabetes etiology, but are poorly understood due to disease heterogeneity, complexity of exposures, and analytical challenges. Machine learning and data mining are artificial intelligence methods that can address these limitations. Despite their increasing adoption in etiology and prediction of diabetes research, the types of methods and exposures analyzed have not been thoroughly reviewed.

OBJECTIVE

We aimed to review articles that implemented machine learning and data mining methods to understand environmental exposures in diabetes etiology and disease prediction.

METHODS

We queried PubMed and Scopus databases for machine learning and data mining studies that used environmental exposures to understand diabetes etiology on September 19th, 2022. Exposures were classified into specific external, general external, or internal exposures. We reviewed machine learning and data mining methods and characterized the scope of environmental exposures studied in the etiology of general diabetes, type 1 diabetes, type 2 diabetes, and other types of diabetes.

RESULTS

We identified 44 articles for inclusion. Specific external exposures were the most common exposures studied, and supervised models were the most common methods used. Well-established specific external exposures of low physical activity, high cholesterol, and high triglycerides were predictive of general diabetes, type 2 diabetes, and prediabetes, while novel metabolic and gut microbiome biomarkers were implicated in type 1 diabetes.

DISCUSSION

The use of machine learning and data mining methods to elucidate environmental triggers of diabetes was largely limited to well-established risk factors identified using easily explainable and interpretable models. Future studies should seek to leverage machine learning and data mining to explore the temporality and co-occurrence of multiple exposures and further evaluate the role of general external and internal exposures in diabetes etiology.

Role of ADMA in the pathogenesis of microvascular complications in type 2 diabetes mellitus

Diabetic microangiopathy is a typical and severe problem in diabetics, including diabetic retinopathy, diabetic nephropathy, diabetic neuropathy, and diabetic cardiomyopathy. Patients with type 2 diabetes and diabetic microvascular complications have significantly elevated levels of Asymmetric dimethylarginine (ADMA), which is an endogenous inhibitor of nitric oxide synthase (NOS). ADMA facilitates the occurrence and progression of microvascular complications in type 2 diabetes through its effects on endothelial cell function, oxidative stress damage, inflammation, and fibrosis. This paper reviews the association between ADMA and microvascular complications of diabetes and elucidates the underlying mechanisms by which ADMA contributes to these complications. It provides a new idea and method for the prevention and treatment of microvascular complications in type 2 diabetes.

A prediction model to assess the risk of egfr loss in patients with type 2 diabetes and preserved kidney function: The amd annals initiative

OBJECTIVE

To develop and validate a model for predicting 5-year eGFR-loss in type 2 diabetes mellitus (T2DM) patients with preserved renal function at baseline.

RESEARCH DESIGN AND METHODS

A cohort of 504.532 T2DM outpatients participating to the Medical Associations of Diabetologists (AMD) Annals Initiative was splitted into the Learning and Validation cohorts, in which the predictive model was respectively developed and validated. A multivariate Cox proportional hazard regression model including all baseline characteristics was performed to identify predictors of eGFR-loss. A weight derived from regression coefficients was assigned to each variable and the overall sum of weights determined the 0 to 8-risk score.

RESULTS

A set of demographic, clinical and laboratory parameters entered the final model. The eGFR-loss score showed a good performance in the Validation cohort. Increasing score values progressively identified a higher risk of GFR loss: a score ≥ 8 was associated with a HR of 13.48 (12.96-14.01) in the Learning and a HR of 13.45 (12.93-13.99) in the Validation cohort. The 5 years-probability of developing the study outcome was 55.9% higher in subjects with a score ≥ 8.

CONCLUSIONS

In the large AMD Annals Initiative cohort, we developed and validated an eGFR-loss prediction model to identify T2DM patients at risk of developing clinically meaningful renal complications within a 5-years time frame.

Pharmacological Treatment of Diabetic and Non-Diabetic Patients With Coronary Artery Disease in the Real World of General Practice

Type 2 diabetes mellitus (T2DM) severely increases the probability of developing coronary artery disease (CAD), and diabetic patients with CAD should be considered at very high cardiovascular risk. The complexity of this clinical scenario makes very hard the appropriateness of the pharmacological treatment in the real world. To investigate the implementation of guideline recommendations for the treatment of patients affected by CAD with or without T2DM, a retrospective observational study was carried out between 2018 and 2020, by using the computerized clinical medical record of 10 general practitioners (GPs) including 13,206 subjects. A total of 926 patients (7.0%) were affected by CAD and 393 (42.4%) of them were also diabetic. LDLc, SBP, DBP, and FPG were recorded in 77.4%, 65.4%, 66.5%, and 82.6% of patients, respectively. Comorbidities (median; IQR = 8; 6–10 vs. 5; 3–7: p < 0.001) were significantly high in diabetic patients. Specialist counselling has been observed in 59.9% of diabetic and 57% of non-diabetic patients (p = 0.400). Antithrombotic drugs, statins, β-blockers, or RAASs were prescribed in 67.2%, 59.6%, and 75.9% of patients, respectively. Overall, 462 (49.9%) patients used the treatment suggested by guidelines. Dyslipidemia, hypertension, atherosclerosis, and specialist counselling were predictors of suggested drugs use both in diabetic and non-diabetic patients. Diabetes was not an independent factor related to the likelihood to be properly treated, according to the guidelines. Glucose lowering drugs were prescribed in 69.5% of diabetic patients, but only 39 (14.3%) were treated with the proper GLP-1 or SGLT2-i, whereas 45 patients (16.5%) received the improper sulphonylureas. Our results showed that a “non-ideal” therapeutic approach was adopted in patients affected by diabetes and CAD. ADA and ESC guidelines recommend the use of at least one hypoglycemic agent belonging to the GLP-1 or SGLT2-i class in diabetic patients with high/very high cardiovascular risk, regardless of the glycemic target (HbA1c <7%). However, only a few diabetic patients on hypoglycemic therapy were appropriately treated. These data suggest that a closer collaboration between the GPs, clinical pharmacologist, and specialists is needed in the real world scenario of the general practice in order to effectively improve adherence to guidelines and overall management of global cardiovascular risk in diabetic patients.