Risk stratification for 25-year cardiovascular disease incidence in type 1 diabetes: Tree-structured survival analysis of the Pittsburgh Epidemiology of Diabetes Complications study

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.

Hyperlipidemia and Cardiovascular Disease in People with Type 1 Diabetes: Review of Current Guidelines and Evidence

PURPOSE OF REVIEW

In this review, we discuss the prevalence of cardiovascular disease in people with type 1 diabetes. We outline key risk factors associated with increased cardiovascular event rates and discuss the prevalence and mechanisms underlying hyperlipidemia in people with type 1 diabetes. Finally, we summarize the evidence to support early and more aggressive lipid-lowering therapy in people with type 1 diabetes and review current guideline recommendations.

RECENT FINDINGS

Comprehensive treatment of hyperglycemia, hypertension, and hyperlipidemia reduces adverse cardiovascular outcomes in people with type 2 diabetes. In contrast, evidence to support a comparable benefit of intensive cardiovascular risk factor management in people with type 1 diabetes is lacking from prospective, randomized trials and has only been shown in registries. Therefore, current treatment guidelines extrapolate prospective clinical trial evidence obtained in people with type 2 diabetes to provide similar treatment recommendations for people with type 1 and type 2 diabetes. Evidence supports the more aggressive treatment of cardiovascular risk factors in people with type 1 diabetes, who would likely benefit from early risk stratification and comprehensive risk factor management, including aggressive lipid-lowering therapy.

Cardiovascular Disease in Adults with Type 1 Diabetes: Looking Beyond Glycemic Control

PURPOSE OF REVIEW

Despite improvements in treatment, people with type 1 diabetes continue to have increased cardiovascular disease (CVD) risk. Glycemic control does not fully explain this excess CVD risk, so a greater understanding of other risk factors is needed.

RECENT FINDINGS

The authors review the relationship between glycemia and CVD risk in adults with type 1 diabetes and summarize evidence regarding other factors that may explain risk beyond glycemia. Insulin resistance, weight gain, sex differences, genetics, inflammation, emerging markers of risk, including lipid subclasses and epigenetic modifications, and future directions are discussed. As glycemic control improves, an increased focus on other CVD risk factors is warranted in type 1 diabetes. Novel markers and precision medicine approaches may improve CVD prediction, but a lack of type 1 diabetes-specific guidelines for lipids, blood pressure, and physical activity are likely impediments to optimal CVD prevention in this high-risk population.

Arterial stiffness in patients with type 1 diabetes and its comparison to cardiovascular risk evaluation tools

Background

Arterial stiffness is a potential biomarker for cardiovascular disease (CVD) risk in patients with type 1 diabetes (T1D). However, its relation with other CV risk evaluation tools in T1D has not been elucidated yet. This study aimed to evaluate arterial stiffness in T1D patients free from known CVD, and compare it to other CV risk evaluation tools used in T1D.

Methods

Cross-sectional study in adults with a T1D duration of at least 10 years and without established CVD. Patients were categorized in CVD risk groups based on 2019 European Society of Cardiology (ESC) guidelines, and the STENO T1D risk engine was used to estimate 10-year risk for CV events. Arterial stiffness was evaluated with carotid-femoral pulse wave velocity (cf-PWV). Coronary artery calcium (CAC) score was assessed and carotid ultrasound was performed. Ambulatory 24-h blood pressure and central hemodynamic parameters were evaluated. Data on renal function and diabetic kidney disease was retrieved.

Results

54 patients (age: 46 ± 9.5 years; T1D duration: 27 ± 8.8 years) were included. One-fourth of patients showed prematurely increased aortic stiffness based on cf-PWV (24%). Cf-PWV was significantly associated with CAC score, carotid intima-media thickness, central hemodynamic parameters and diabetic kidney disease. Based on STENO, 20 patients (37%) were at low, 20 patients (37%) at moderate, and 14 patients (26%) at high 10-year risk for CV event. Cf-PWV was strongly associated with the STENO score (r_s = + 0.81; R² = 0.566, p < 0.001), increasing with each higher STENO group (p < 0.01). However, cf-PWV was not significantly different between the two CV risk groups (high versus very high) based on ESC criteria, and ESC criteria compared to STENO classified 10 patients more as having > 10% 10-year risk for CV events (n = 44/54; 81.5% versus n = 34/54; 63%).

Conclusions

This study demonstrated that a substantial proportion of long-standing T1D patients free from known CVD show premature arterial stiffening. Cf-PWV strongly associates with the STENO risk score for future CV events and with cardiovascular imaging and function outcomes, thereby illustrating the clinical importance of arterial stiffness. The data, however, also show considerable heterogeneity in CV risk and differences in risk categorisation between the STENO tool and ESC criteria.There is a need for refinement of CV risk classification in T1D, and future studies should investigate if evaluation of arterial stiffness should be implemented in T1D clinical practice and which patients benefit the most from its assessment.

Risk factors differ by first manifestation of cardiovascular disease in type 1 diabetes

AIMS

We compared risk factors for three CVD manifestations and a composite outcome over 25 years' follow-up in the Pittsburgh Epidemiology of Diabetes Complications (EDC) prospective cohort study of childhood-onset (<17 years) type 1 diabetes (n = 658).

METHODS

First CVD manifestations examined were: (1) major atherosclerotic cardiovascular event (MACE, i.e. CVD death, myocardial infarction, stroke), (2) coronary revascularization, (3) soft coronary artery disease (CAD, i.e. ischemia ECG, angina), and a (4) composite (MACE + revascularization) outcome. Baseline and time-varying mean and current risk factors, including medication use, were assessed, in diabetes duration-adjusted models.

RESULTS

MACE (n = 107) was predicted by ln(albumin excretion rate) (AER, HR = 1.3, p < 0.0001), systolic BP (SBP, HR = 1.03, p < 0.0001), white blood cell count (WBC, HR = 1.2, p < 0.0001), HbA1c (HR = 1.2p = 0.03), LDLc (HR = 1.01, p = 0.03). Soft CAD (n = 91) was predicted by ln(AER) (HR = 1.2, p = 0.004), SBP (HR = 1.03, p = 0.0002), WBC (HR = 1.2, p = 0.0003), HbA1c (HR = 1.2, p = 0.005). Revascularization (n = 38) was predicted by LDLc (HR = 1.03, p < 0.0001), eGFR (HR = 0.98, p = 0.002), HbA1c (HR = 1.3, p = 0.03). Adding revascularization to MACE enhanced the role of LDLc, while diminishing that of HbA1c, compared to MACE alone.

CONCLUSIONS

Important risk factor associations may be affected by examining composite CVD outcomes. More research is needed to determine how to best incorporate revascularization into composite CVD definitions.

Risk Factor Modeling for Cardiovascular Disease in Type 1 Diabetes in the Pittsburgh Epidemiology of Diabetes Complications (EDC) Study: A Comparison With the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Study (DCCT/EDIC)

In a recent Diabetes Control and Complications Trial (DCCT)/Epidemiology of Diabetes Interventions and Complications (EDIC) study report, mean HbA_1c was the strongest predictor of cardiovascular disease (CVD) after age. In DCCT/EDIC, mean diabetes duration was 6 years (median 4) at baseline and those with high blood pressure or cholesterol were excluded. We now replicate these analyses in the Pittsburgh Epidemiology of Diabetes Complications (EDC) prospective cohort study of childhood-onset (at <17 years of age) type 1 diabetes, with similar age (mean 27 years in both studies) but longer diabetes duration (mean 19 years and median 18 years) and no CVD risk factor exclusion at baseline. CVD incidence (CVD death, myocardial infarction (MI), stroke, revascularization, angina, or ischemic electrocardiogram) was associated with diabetes duration, most recent albumin excretion rate (AER), updated mean triglycerides, baseline hypertension, baseline LDL cholesterol, and most recent HbA_1c Major atherosclerotic cardiovascular events (CVD death, MI, or stroke) were associated with diabetes duration, most recent AER, baseline systolic blood pressure, baseline smoking, and updated mean HbA_1c Compared with findings in DCCT/EDIC, traditional risk factors similarly predicted CVD; however AER predominates in EDC and HbA_1c in DCCT/EDIC. Thus, the relative impact of HbA_1c and kidney disease in type 1 diabetes varies according to diabetes duration.

Artificial Intelligence Methodologies and Their Application to Diabetes

In the past decade diabetes management has been transformed by the addition of continuous glucose monitoring and insulin pump data. More recently, a wide variety of functions and physiologic variables, such as heart rate, hours of sleep, number of steps walked and movement, have been available through wristbands or watches. New data, hydration, geolocation, and barometric pressure, among others, will be incorporated in the future. All these parameters, when analyzed, can be helpful for patients and doctors' decision support. Similar new scenarios have appeared in most medical fields, in such a way that in recent years, there has been an increased interest in the development and application of the methods of artificial intelligence (AI) to decision support and knowledge acquisition. Multidisciplinary research teams integrated by computer engineers and doctors are more and more frequent, mirroring the need of cooperation in this new topic. AI, as a science, can be defined as the ability to make computers do things that would require intelligence if done by humans. Increasingly, diabetes-related journals have been incorporating publications focused on AI tools applied to diabetes. In summary, diabetes management scenarios have suffered a deep transformation that forces diabetologists to incorporate skills from new areas. This recently needed knowledge includes AI tools, which have become part of the diabetes health care. The aim of this article is to explain in an easy and plane way the most used AI methodologies to promote the implication of health care providers-doctors and nurses-in this field.

Whole Egg Consumption Exerts a Nephroprotective Effect in an Acute Rodent Model of Type 1 Diabetes

Nephropathy is a well-characterized complication of type 1 diabetes (T1D), resulting in proteinuria and urinary loss of micronutrients. We previously found that a whole egg-based diet maintained vitamin D balance in type 2 diabetic rats despite excessive urinary losses due to nephropathy. The goal of this study was to investigate the impact of whole egg consumption in T1D rats. Sprague-Dawley rats were randomly assigned to T1D or nondiabetic control groups and fed a casein or whole egg-based diet for 32 days. On day 26, two-thirds of the rats received a streptozotocin injection to induce T1D. Whole egg consumption attenuated polyuria, proteinuria, and renal hypertrophy in T1D rats. These data suggest that dietary intervention with whole egg may offer renal protection in T1D.

Cardiovascular complications of type 1 diabetes: update on the renal link

AIMS

Despite recent findings of increased life expectancy among individuals with type 1 diabetes, mortality remains greatly increased compared to the general population. As this is largely the result of cardiovascular and renal complications, we aimed to review recent findings surrounding these diseases in type 1 diabetes.

METHODS

We reviewed published findings concerning the cardiovascular complications of type 1 diabetes, with a particular focus on links with renal disease.

RESULTS

The cardiovascular and renal complications of type 1 diabetes share many features including insulin resistance, oxidative damage, and genetic associations with the Haptoglobin genotype, and both are strongly affected by glycemic control.

CONCLUSIONS

Although current knowledge on predictors of type 1 diabetes cardiovascular and renal complications has increased, further investigation is required to understand the mechanisms leading to cardio-renal complications in this population.