Association of prediabetes and diabetes with stroke symptoms: the REasons for Geographic and Racial Differences in Stroke (REGARDS) study

Machine Learning Approaches for Stroke Risk Prediction: Findings from the Suita Study

Stroke constitutes a significant public health concern due to its impact on mortality and morbidity. This study investigates the utility of machine learning algorithms in predicting stroke and identifying key risk factors using data from the Suita study, comprising 7389 participants and 53 variables. Initially, unsupervised k-prototype clustering categorized participants into risk clusters, while five supervised models including Logistic Regression (LR), Random Forest (RF), Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), and Light Gradient Boosted Machine (LightGBM) were employed to predict stroke outcomes. Stroke incidence disparities among identified risk clusters using the unsupervised k-prototype clustering method are substantial, according to the findings. Supervised learning, particularly RF, was a preferable option because of the higher levels of performance metrics. The Shapley Additive Explanations (SHAP) method identified age, systolic blood pressure, hypertension, estimated glomerular filtration rate, metabolic syndrome, and blood glucose level as key predictors of stroke, aligning with findings from the unsupervised clustering approach in high-risk groups. Additionally, previously unidentified risk factors such as elbow joint thickness, fructosamine, hemoglobin, and calcium level demonstrate potential for stroke prediction. In conclusion, machine learning facilitated accurate stroke risk predictions and highlighted potential biomarkers, offering a data-driven framework for risk assessment and biomarker discovery.

Identifying diagnostic indicators for type 2 diabetes mellitus from physical examination using interpretable machine learning approach

Background

Identification of patients at risk for type 2 diabetes mellitus (T2DM) can not only prevent complications and reduce suffering but also ease the health care burden. While routine physical examination can provide useful information for diagnosis, manual exploration of routine physical examination records is not feasible due to the high prevalence of T2DM.

Objectives

We aim to build interpretable machine learning models for T2DM diagnosis and uncover important diagnostic indicators from physical examination, including age- and sex-related indicators.

Methods

In this study, we present three weighted diversity density (WDD)-based algorithms for T2DM screening that use physical examination indicators, the algorithms are highly transparent and interpretable, two of which are missing value tolerant algorithms.

Patients

Regarding the dataset, we collected 43 physical examination indicator data from 11,071 cases of T2DM patients and 126,622 healthy controls at the Affiliated Hospital of Southwest Medical University. After data processing, we used a data matrix containing 16004 EHRs and 43 clinical indicators for modelling.

Results

The indicators were ranked according to their model weights, and the top 25% of indicators were found to be directly or indirectly related to T2DM. We further investigated the clinical characteristics of different age and sex groups, and found that the algorithms can detect relevant indicators specific to these groups. The algorithms performed well in T2DM screening, with the highest area under the receiver operating characteristic curve (AUC) reaching 0.9185.

Conclusion

This work utilized the interpretable WDD-based algorithms to construct T2DM diagnostic models based on physical examination indicators. By modeling data grouped by age and sex, we identified several predictive markers related to age and sex, uncovering characteristic differences among various groups of T2DM patients.

Atherosclerotic cardiovascular disease events among adults with high predicted risk without established risk factors

Objective

Age is the strongest contributor to 10-year predicted atherosclerotic cardiovascular disease (ASCVD) risk. Some older adults have a predicted ASCVD risk ≥7.5 %, without established risk factors. We sought to compare ASCVD incidence among adults with predicted ASCVD risk ≥7.5 %, with and without established ASCVD risk factors, to adults with predicted risk <7.5 %.

Methods

We analyzed data from REasons for Geographic and Racial Differences in Stroke study participants, 45-79 years old, without ASCVD or diabetes, not taking statins and with low-density lipoprotein cholesterol 70-189 mg/dL. Participants were categorized into 3 groups based on their 10-year predicted ASCVD risk and presence of established risk factors: <7.5 %, ≥7.5 % with established risk factors and ≥7.5 % without established risk factors. Established risk factors included smoking, systolic blood pressure ≥130 mmHg or antihypertensive medication use, total cholesterol ≥200 mg/dL, or high-density lipoprotein cholesterol <50 mg/dL for women (<40 mg/dL for men). Participants were followed for ASCVD events.

Results

Among 11,115 participants, 911 incident ASCVD events occurred over a median of 11.1 years. ASCVD incidence rates were 3.6, 12.8, and 9.8 per 1,000 person-years for participants with predicted risk <7.5 %, predicted risk ≥7.5 % with established risk factors and predicted risk ≥7.5 % without established risk factors, respectively. Compared to adults with predicted risk <7.5 %, hazard ratios for incident ASCVD in participants with risk ≥7.5 % with and without established risk factors were 3.58 (95 %CI 3.03 - 4.21) and 2.72 (95 %CI 1.91-3.88), respectively.

Conclusions

Adults with a 10-year predicted ASCVD risk ≥7.5 % but without established risk factors had a high ASCVD incidence.

D-dimer and the risk of hypertension: The REasons for Geographic And Racial Differences in Stroke Cohort Study

Background

Reasons for increased risk of hypertension in Black compared with White people are only partly understood. D-dimer, a thrombo-inflammatory marker higher in Black individuals, is also higher in people with hypertension. However, the impact of D-dimer on racial disparities in risk of incident hypertension has not been studied.

Objectives

To assess whether D-dimer is associated with the risk of incident hypertension, whether the association between D-dimer and the risk of incident hypertension differs by race, and whether the biology reflected by D-dimer explains racial disparities in the risk of incident hypertension.

Methods

This study included 1867 participants in the REasons for Geographic And Racial Differences in Stroke cohort study without baseline hypertension and with a second visit 9.4 years after baseline. Risk ratios of incident hypertension by baseline D-dimer level were estimated, a D-dimer-by-race interaction was tested, and the mediating effect of D-dimer (which represents underlying biological processes) on the association of race and hypertension risk was assessed.

Results

The risk of incident hypertension was 47% higher in persons in the top quartile than in those in the bottom quartile of D-dimer (risk ratio [RR]: 1.47; 95% CI: 1.23-1.76). The association was partly attenuated after adjusting for sociodemographic and adiposity-related risk factors (RR: 1.22; 95% CI: 1.02-1.47). The association of D-dimer and hypertension did not differ by race, and D-dimer did not attenuate the racial difference in the risk of incident hypertension.

Conclusion

D-dimer concentration reflects pathophysiology related to the development of hypertension. Specific mechanisms require further study and may involve adiposity.

Elevated depressive symptoms and risk of all-cause and cardiovascular mortality among adults with and without diabetes: The REasons for Geographic And Racial Differences in Stroke (REGARDS) study

AIMS

To examine the association of elevated depressive symptoms with all-cause and cardiovascular disease (CVD) mortality and determine whether these associations differ for those with and without diabetes.

METHODS

We included 22,807 black and white men and women aged 45-98 years at baseline (2003-2007) from the REasons for Geographic And Racial Differences in Stroke (REGARDS) Study. Elevated depressive symptoms were defined as a score ≥ 4 on the 4-item Centers for Epidemiologic Studies of Depression Scale. Participants were classified as having diabetes, prediabetes, or no prediabetes/diabetes based on glucose levels and diabetes medication use. All-cause mortality events were available through 2018 and adjudicated CVD mortality events were available through 2015.

RESULTS

During follow-up, there were 5383 all-cause deaths, of which 1585 were adjudicated CVD deaths. The mean survival time was lower for participants with elevated depressive symptoms than those without elevated depressive symptoms for those with diabetes, prediabetes, and no prediabetes/diabetes. In multivariable adjusted models, elevated depressive symptoms increased the risk of all-cause mortality for those with diabetes (HR = 1.15; 95% CI = 1.00-1.32), prediabetes (HR = 1.56; 95% CI = 1.28-1.91), and neither prediabetes/diabetes (HR = 1.34; 95% CI = 1.19-1.50) (p for interaction = 0.0342). Findings were similar for CVD mortality.

CONCLUSION

Elevated depressive symptoms increased the risk of all-cause and CVD mortality among individuals with and without diabetes, with a stronger magnitude of association observed among those with prediabetes. This underscores the need for assessing depressive symptoms across the glycemic spectrum, including those with prediabetes.

Examining the causal mediating role of brain pathology on the relationship between diabetes and cognitive impairment: the Cardiovascular Health Study

The paper examines whether diabetes mellitus leads to incident mild cognitive impairment and dementia through brain hypoperfusion and white matter disease. We performed inverse odds ratio weighted causal mediation analyses to decompose the effect of diabetes on cognitive impairment into direct and indirect effects, and we found that approximately a third of the total effect of diabetes is mediated through vascular-related brain pathology. Our findings lend support for a common aetiological hypothesis regarding incident cognitive impairment, which is that diabetes increases the risk of clinical cognitive impairment in part by impacting the vasculature of the brain.

Does the Association of Diabetes With Stroke Risk Differ by Age, Race, and Sex? Results From the REasons for Geographic and Racial Differences in Stroke (REGARDS) Study

OBJECTIVE

Given temporal changes in diabetes prevalence and stroke incidence, this study investigated age, race, and sex differences in the diabetes-stroke association in a contemporary prospective cohort, the REasons for Geographic and Racial Differences in Stroke (REGARDS) Study.

RESEARCH DESIGN AND METHODS

We included 23,002 non-Hispanic black and white U.S. adults aged ≥45 years without prevalent stroke at baseline (2003-2007). Diabetes was defined as fasting glucose ≥126 mg/dL, random glucose ≥200 mg/dL, or use of glucose-lowering medication. Incident stroke events were expert adjudicated and available through September 2017.

RESULTS

The prevalence of diabetes was 19.1% at baseline. During follow-up, 1,018 stroke events occurred. Among adults aged <65 years, comparing those with diabetes to those without diabetes, the risk of stroke was increased for white women (hazard ratio [HR] 3.72 [95% CI 2.10-6.57]), black women (HR 1.88 [95% CI 1.22-2.90]), and white men (HR 2.01 [95% CI 1.27-3.27]) but not black men (HR 1.27 [95% CI 0.77-2.10]) after multivariable adjustment. Among those aged ≥65 years, diabetes increased the risk of stroke for white women and black men, but not black women (HR 1.05 [95% CI 0.74-1.48]) or white men (HR 0.86 [95% CI 0.62-1.21]).

CONCLUSIONS

In this contemporary cohort, the diabetes-stroke association varied by age, race, and sex together, with a more pronounced effect observed among adults aged <65 years. With the recent increase in the burden of diabetes complications at younger ages in the U.S., additional efforts are needed earlier in life for stroke prevention among adults with diabetes.

A PheWAS study of a large observational epidemiological cohort of African Americans from the REGARDS study

BACKGROUND

Cardiovascular disease, diabetes, and kidney disease are among the leading causes of death and disability worldwide. However, knowledge of genetic determinants of those diseases in African Americans remains limited.

RESULTS

In our study, associations between 4956 GWAS catalog reported SNPs and 67 traits were examined among 7726 African Americans from the REasons for Geographic and Racial Differences in Stroke (REGARDS) study, which is focused on identifying factors that increase stroke risk. The prevalent and incident phenotypes studied included inflammation, kidney traits, cardiovascular traits and cognition. Our results validated 29 known associations, of which eight associations were reported for the first time in African Americans.

CONCLUSION

Our cross-racial validation of GWAS findings provide additional evidence for the important roles of these loci in the disease process and may help identify genes especially important for future functional validation.

Interaction of Sex and Diabetes on Outcome After Ischemic Stroke

Background

The relationship between ischemic stroke (IS), diabetes mellitus (DM), and sex is intriguing. The aim of this study was to assess the effect modification of sex in the association between DM and short- and long-term disability and mortality in first-ever IS patients.

Methods

In a retrospective, observational, hospital-based study of a prospective series including first-ever IS patients from January 2006 until July 2011, differences in 3-month and 5-year mortality, and disability between diabetic and non-diabetic patients [modified Rankin Scale (mRS) from 3 to 5] were analyzed by sex.

Results

In total, 933 patients (36.3% with DM, 50.5% women) were included. Overall 3-month and 5-year mortality were 150 (16.1%) and 407 (44.1%), respectively. Adjusted for age, previous mRS, and stroke severity, patients with DM had significantly higher 3-month disability [hazard ratio (HR): 1.49 (95% confidence interval (CI): 1.39–1.70), p < 0.0001], 5-year disability [HR: 1.41 (95% CI: 1.07–1.86), p = 0.015], and 5-year mortality [HR: 1.48 (95% CI: 1.20–1.81), p < 0.0001], compared with the non-DM group. Compared with non-DM women, women with diabetes had worse 3-month disability [HR: 1.81 (95% CI: 1.33–2.46), p < 0.0001] and 5-year mortality [HR: 1.72 (95% CI: 1.30–2.20), p < 0.0001], and a trend for 5-year disability [HR: 1.40 (95% CI: 0.99–2.09), p = 0.057]. In men, DM had an effect on 3-month disability [HR: 1.45 (95% CI: 1.07–1.96), p = 0.018], a trend for 5-year disability [HR: 1.43 (95% CI: 0.94–2.19), p = 0.096], but no clear effect on 5-year mortality [HR: 1.22 (95% CI: 0.91–1.65), p = 0.186].

Conclusion

Sex has a modifier effect on mortality in first-ever IS diabetic patients. Long-term mortality is increased in diabetic women compared with non-diabetic women, a difference not observed in men.

Determinants of Stroke in Brazil: A Cross-Sectional Multivariate Approach from the National Health Survey

OBJECTIVE

The objective of this study was to investigate the association between biological-health, socioeconomics, and behavioral determinants and stroke by evaluating the risk of stroke in the presence of each or all determinants grouped in a multivariate model.

METHODS

This study is a transversal secondary data analysis of the Brazilian National Health Survey, released by the Brazilian Institute of Geography and Statistics. The prevalence, prevalence ratios, and confidence intervals were calculated for the main outcome. A multivariate regression model was applied, with the stroke as outcome and all other variables with a P value of .20 or lower in the univariate analysis included as explanatory variables to adjust for potential confounders and effect modifiers.

RESULTS

The mean age was 43.3 ± 16.6 years, ranging from 18 to 101 years. The prevalence of hypertension was 21.4%, and with regard to lifestyle habits, it was observed that 20.0% had smoked but stopped and 29.7% practiced physical activity in the last 30 days. The regression model showed that the odds ratio in the final model was weighted, with low schooling, smoking habit, overweight, low physical activity practice, diabetes, and hypertension being significantly associated with stroke.

CONCLUSION

The multivariate model showed that the chance of stroke is high, both combined or isolated.

Diabetes, diabetes severity, and coronary heart disease risk equivalence: REasons for Geographic and Racial Differences in Stroke (REGARDS)

BACKGROUND

Evidence is mixed regarding whether diabetes confers equivalent risk of coronary heart disease (CHD) as prevalent CHD. We investigated whether diabetes and severe diabetes are CHD risk equivalents.

METHODS

At baseline, participants in the REasons for Geographic and Racial Differences in Stroke (REGARDS) study (black and white US adults ≥45 years old recruited in 2003-2007) were categorized as having prevalent CHD only (self-reported or electrocardiogram evidence; n = 3,043), diabetes only (self-reported or elevated glucose; n = 4,012), diabetes and prevalent CHD (n = 1,529), and neither diabetes nor prevalent CHD (n = 17,155). Participants with diabetes using insulin and/or with albuminuria (urinary albumin-to-creatinine ratio ≥30 mg/g) were categorized as having severe diabetes. Participants were followed up through 2011 for CHD events (myocardial infarction or fatal CHD).

RESULTS

During a mean follow-up of 5 years, 1,385 CHD events occurred. The hazard ratios of CHD events comparing participants with diabetes only, diabetes, and prevalent CHD and neither diabetes nor prevalent CHD with those with prevalent CHD were 0.65 (95% CI 0.54-0.77), 1.54 (95% CI 1.30-1.83), and 0.41 (95% CI 0.35-0.47), respectively, after adjustment for demographics and risk factors. Compared with participants with prevalent CHD, the hazard ratio of CHD events for participants with severe diabetes was 0.88 (95% CI 0.72-1.09).

CONCLUSIONS

Participants with diabetes had lower risk of CHD events than did those with prevalent CHD. However, participants with severe diabetes had similar risk to those with prevalent CHD. Diabetes severity may need consideration when deciding whether diabetes is a CHD risk equivalent.

Stroke symptoms and risk for incident coronary heart disease in the REasons for Geographic And Racial Differences in Stroke (REGARDS) study

BACKGROUND

Many adults without cerebrovascular disease report a history of stroke symptoms, which is associated with higher risk for stroke. Because stroke and coronary heart disease (CHD) share many risk factors, we examined the association between a history of stroke symptoms and incident CHD.

METHODS

We analyzed data from 8999 black and 12,499 white REasons for Geographic And Racial Differences in Stroke (REGARDS) study participants without a prior myocardial infarction, stroke or transitory ischemic attack enrolled in 2003-2007 (total participants=21,498, all ≥45years of age). A history of stroke symptoms (i.e., unilateral weakness, unilateral numbness, full-field vision loss, half-field vision loss, understanding problems and communication problems) was assessed at baseline using the Questionnaire for Verifying Stroke-Free Status. Participants were followed for incident CHD and CHD death through December 2011.

RESULTS

Overall, 3432 (16.0%) participants reported a history of stroke symptoms (1771 [19.7%] blacks and 1661 [13.3%] whites). There were 701 incident CHD events including 209 CHD deaths over a median follow-up of 5.8years. After adjustment for CHD risk factors, hazard ratios (95% confidence interval [95% CI]) for incident CHD associated with reporting any versus no stroke symptoms were 1.26 (1.04-1.51) in the overall population, 1.28 (0.99-1.65) among blacks and 1.23 (0.94-1.61) among whites. Multivariable-adjusted hazard ratios (95% CI) for CHD death associated with any versus no stroke symptoms were 1.50 (1.10-2.06) overall, 1.58 (1.07-2.32) among blacks and 1.41 (0.82-2.43) among whites.

CONCLUSION

A history of stroke symptoms is associated with a higher incidence of CHD among black and white adults.

Subclinical hypothyroidism and the associations with macrovascular complications and chronic kidney disease in patients with Type 2 diabetes

AIMS

The prevalence of subclinical hypothyroidism (SCH) is high among patients with diabetes, although the relationship between SCH and diabetic vascular complications is unknown. This study aimed to determine the relationship between SCH and vascular complications in patients with Type 2 diabetes.

METHODS

In this cross sectional study, 991 patients with Type 2 diabetes were screened for thyroid function at their admission to the Second Affiliated Hospital of Chongqing Medical University. We compared the prevalence of coronary heart disease (CHD), ischaemic stroke and chronic kidney disease (CKD) with the prevalence of euthyroidism and SCH.

RESULTS

Among the 991 patients, 126 (12.7%) patients had SCH. The prevalence of CHD was significantly higher in the SCH group than in the euthyroid group (22.2% and 15.0%, respectively; P = 0.039). In the logistic regression analyses, SCH was associated with CHD [odds ratio (OR): 1.993; 95% confidence interval (CI): 1.135-3.497; P = 0.016]. This association was stronger in patients aged ≥ 65 years than in younger patients [2.474 (1.173-5.220); P = 0.017]. No significant association was found between SCH and ischaemic stroke. Patients with severe SCH had a high risk of CKD [1.842 (1.120-3.029); P = 0.016].

CONCLUSIONS

This study provides evidence that SCH in patients with Type 2 diabetes is associated with a high prevalence of CHD (and CKD in severe SCH), although not with ischaemic stroke.

The role of HbA1c determination in detecting unknown glucose disturbances in ischemic stroke

Objectives

To evaluate the usefulness of hemoglobin A1c (HbA1c) determinations during the acute ischemic stroke (IS) to identify undiagnosed glucose disturbances in a prospective series of patients with first-ever IS.

Methods

Retrospective analysis of a prospective series of first-ever IS patients. Patients with previous diagnosis of diabetes mellitus (DM) were excluded from the study. Patients were classified as non-DM (HbA1c<5.7% and no previous evidence of 2 or more fasting blood glucose> = 126 mg/dL), prediabetes (HbA1c from 5.7% to 6.4%), and new suspected DM (HbA1c> = 6.5% independently of current blood glucose). Medical charts from hospital discharge to July 2014 of all suspected DM patients were reviewed to confirm the DM diagnosis.

Results

The initial cohort included 1283 patients, of which 393 were excluded because of previous DM diagnosis and 136 because HbA1c during acute stroke phase was not available. No demographic differences were observed between patients with and without HbA1c determinations. The final cohort was composed of 754 patients with first-ever IS and unknown DM history. HbA1c determination suggested new DM in 87 cases (11.5%) and detected 273 patients with prediabetes (36.2%). New DM cases were identified in all etiological stroke subtypes. After discharge, DM diagnosis was confirmed in 80.2% of patients with available follow-up.

Conclusions

HbA1c determination detected both undiagnosed DM and prediabetes in IS patients without taking into account the blood glucose values during admission, and independently of etiological stroke subtype. HbA1c determination should be included in the systematic screening of all IS patients.

Ischemic stroke in prediabetic patients

To describe the clinical characteristics of first-ever ischemic stroke (IS) patients with prediabetes, and to compare them with diabetes mellitus (DM) and non-DM patient characteristics. Retrospective analysis of a prospective series of first-ever acute IS patients. Patients were classified as non-DM (HbA1c during admission <5.7% and no previous evidence of 2 or more fasting glucose >126 mg/dL), prediabetes (HbA1c from 5.7 to 6.4%), and DM (previous DM diagnosis or HbA1c ≥6.5% independently of current blood glucose). Demographic and clinical characteristics were compared between the three groups, along with outcome data [early neurological deterioration (END), 3-month poor outcome, 3-month mortality, outcome after rtPA treatment]. No demographic differences were observed. Prediabetic patients had more arterial hypertension (p = 0.006) and higher waist circumference (p < 0.0001) than non-DM patients, and DM patients had more hypercholesterolemia (p < 0.0001), body mass index (p = 0.017), and coronary artery disease (p = 0.005) than prediabetics. There were differences in TOAST subtype distribution (p < 0.0001). There were no differences in rtPA treatment success rate between groups. Multivariate analysis adjusted by age and stroke severity showed that DM but not prediabetes is an independent factor associated with END and 3-month poor outcome. Prediabetic patients with IS exhibit an "intermediate" vascular risk factor profile between that of non-DM and DM patients. In contrast to DM patients, IS prognosis in patients with prediabetes is similar to non-DM patients.