1
|
Wu CZ, Huang LY, Chen FY, Kuo CH, Yeih DF. Using Machine Learning to Predict Abnormal Carotid Intima-Media Thickness in Type 2 Diabetes. Diagnostics (Basel) 2023; 13:diagnostics13111834. [PMID: 37296685 DOI: 10.3390/diagnostics13111834] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/02/2023] [Revised: 05/16/2023] [Accepted: 05/20/2023] [Indexed: 06/12/2023] Open
Abstract
Carotid intima-media thickness (c-IMT) is a reliable risk factor for cardiovascular disease risk in type 2 diabetes (T2D) patients. The present study aimed to compare the effectiveness of different machine learning methods and traditional multiple logistic regression in predicting c-IMT using baseline features and to establish the most significant risk factors in a T2D cohort. We followed up with 924 patients with T2D for four years, with 75% of the participants used for model development. Machine learning methods, including classification and regression tree, random forest, eXtreme gradient boosting, and Naïve Bayes classifier, were used to predict c-IMT. The results showed that all machine learning methods, except for classification and regression tree, were not inferior to multiple logistic regression in predicting c-IMT in terms of higher area under receiver operation curve. The most significant risk factors for c-IMT were age, sex, creatinine, body mass index, diastolic blood pressure, and duration of diabetes, sequentially. Conclusively, machine learning methods could improve the prediction of c-IMT in T2D patients compared to conventional logistic regression models. This could have crucial implications for the early identification and management of cardiovascular disease in T2D patients.
Collapse
Affiliation(s)
- Chung-Ze Wu
- Division of Endocrinology and Metabolism, Department of Internal Medicine, School of Medicine, College of Medicine, Taipei Medical University, Taipei City 11031, Taiwan
- Division of Endocrinology and Metabolism, Department of Internal Medicine, Shuang Ho Hospital, Taipei Medical University, New Taipei City 23561, Taiwan
| | - Li-Ying Huang
- Division of Endocrinology and Metabolism, Department of Internal Medicine, Department of Medical Education, Fu Jen Catholic University Hospital, Fu Jen Catholic University, New Taipei City 24352, Taiwan
- School of Medicine, College of Medicine, Fu Jen Catholic University, New Taipei City 24352, Taiwan
| | - Fang-Yu Chen
- School of Medicine, College of Medicine, Fu Jen Catholic University, New Taipei City 24352, Taiwan
- Division of Endocrinology and Metabolism, Department of Internal Medicine, Fu Jen Catholic University Hospital, Fu Jen Catholic University, New Taipei City 24352, Taiwan
| | - Chun-Heng Kuo
- School of Medicine, College of Medicine, Fu Jen Catholic University, New Taipei City 24352, Taiwan
- Division of Endocrinology and Metabolism, Department of Internal Medicine, Fu Jen Catholic University Hospital, Fu Jen Catholic University, New Taipei City 24352, Taiwan
| | - Dong-Feng Yeih
- School of Medicine, College of Medicine, Fu Jen Catholic University, New Taipei City 24352, Taiwan
- Division of Cardiology, Department of Internal Medicine, Fu Jen Catholic University Hospital, Fu Jen Catholic University, New Taipei City 24352, Taiwan
| |
Collapse
|
2
|
Abstract
Polyunsaturated fatty acids are of particular interest in the nutritional therapy for diabetes, given their potential role in several pathophysiological processes related to cardiovascular disease. Both omega-3 and omega-6 fatty acids are beneficial for improving lipid profiles in healthy individuals and among type 2 diabetic patients: Supplementation with omega-3 fatty acids lowers triglycerides and VLDL-cholesterol. However, they might also increase LDL-cholesterol. Omega-3 fatty acids are, from the latest evidence, not related to mortality and cardiovascular disease. Similarly, glucose control and hypertension, as well as risk of microvascular complications, seem unaffected by omega-3 supplementation. Most studies involved mainly patients with type 2 diabetes, and future research needs to focus on the type 1 diabetic patient. Also, the role of omega-6 fatty acids remains largely unknown.
Collapse
MESH Headings
- Cardiovascular Diseases/blood
- Cardiovascular Diseases/prevention & control
- Cholesterol, LDL/blood
- Cholesterol, VLDL/blood
- Diabetes Mellitus, Type 1/blood
- Diabetes Mellitus, Type 1/diet therapy
- Diabetes Mellitus, Type 1/physiopathology
- Diabetes Mellitus, Type 2/blood
- Diabetes Mellitus, Type 2/diet therapy
- Diabetes Mellitus, Type 2/physiopathology
- Fatty Acids, Omega-3/therapeutic use
- Fatty Acids, Omega-6/therapeutic use
- Humans
- Triglycerides/blood
Collapse
Affiliation(s)
- Charlotte Jeppesen
- Department of Molecular Epidemiology, German Institute of Human Nutrition, Arthur-Scheunert-Allee 114-116, 14558, Nuthetal, Germany.
| | | | | |
Collapse
|
3
|
Associations between blood glucose and carotid intima-media thickness disappear after adjustment for shared risk factors: the KORA F4 study. PLoS One 2012; 7:e52590. [PMID: 23285104 PMCID: PMC3528645 DOI: 10.1371/journal.pone.0052590] [Citation(s) in RCA: 28] [Impact Index Per Article: 2.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/13/2012] [Accepted: 11/20/2012] [Indexed: 12/20/2022] Open
Abstract
Objective The association between blood glucose and carotid intima-media thickness (CIMT) is considered to be established knowledge. We aimed to assess whether associations between different measures of glycaemia and CIMT are actually independent of anthropometric variables and metabolic risk factors. Moreover, we checked published studies for the adjustment for shared risk factors of blood glucose and CIMT. Methods Fasting glucose, 2-hour glucose, HbA1c, and CIMT were measured in 31-81-years-old participants of the population-based Cooperative Health Research in the Region of Augsburg (KORA) F4 study in Southern Germany (n = 2,663). CIMT was assessed according to the Rotterdam protocol. Linear and logistic regression models with adjustment for age, sex, anthropometric measures, hypertension, and dyslipidaemia were fitted to assess the association between continuous measures of glycaemia, and categories of glucose regulation, respectively, with CIMT. Results We found a 0.10 mm increase (95%-confidence interval: 0.08–0.12) in CIMT in subjects with compared to subjects without diabetes in crude analysis. This increase was not significant in age-sex adjusted models (p = 0.17). Likewise, neither impaired fasting glucose (p = 0.22) nor impaired glucose tolerance (p = 0.93) were associated with CIMT after adjustment for age, sex, and waist circumference. In multivariable adjusted models, age, sex, hypertension, waist circumference, HDL and LDL cholesterol, but neither fasting glucose nor 2-hour glucose nor HbA1c were associated with elevated CIMT. Literature findings are inconclusive regarding an independent association of glucose levels and CIMT. Conclusion CIMT is highly dependent on traditional cardiovascular risk factors, but no relationships between blood glucose and CIMT were found after adjustment for age, sex, and anthropometric variables.
Collapse
|
4
|
Association of Vascular Risk Factors With Carotid Intima Media Thickness After Kidney Transplant. Transplantation 2010. [DOI: 10.1097/tp.0b013e3181f62fe4] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/25/2022]
|
5
|
Wakabayashi I, Masuda H. Relationships Between Vascular Indexes and Atherosclerotic Risk Factors in Patients With Type 2 Diabetes Mellitus. Angiology 2008; 59:567-73. [DOI: 10.1177/0003319707312517] [Citation(s) in RCA: 10] [Impact Index Per Article: 0.6] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/04/2023]
Abstract
The relationships of cardio-ankle vascular index, (a new indicator of arterial stiffness), common carotid artery intima-media thickness, and ankle-brachial pressure index with atherosclerotic risk factors were compared in patients with type 2 diabetes. There were significant correlations between each pair of the above vascular indexes. Both cardio-ankle vascular index and intima-media thickness showed significant correlations with age, systolic blood pressure, pulse pressure, and serum total cholesterol. Ankle-brachial pressure index showed a significant negative correlation with pulse pressure but not with systolic blood pressure and total cholesterol. Logistic regression analysis showed that cardio-ankle vascular index and intima-media thickness were associated with systolic blood pressure and pulse pressure independently of age. These results suggest that there are significant associations among the 3 vascular indexes and that systolic blood pressure and pulse pressure are major, age-independent determinants of cardio-ankle vascular index and common carotid artery intima-media thickness in patients with diabetes.
Collapse
Affiliation(s)
- Ichiro Wakabayashi
- Department of Environmental and Preventive Medicine, Hyogo College of Medicine,
| | - Hiroshi Masuda
- Department of Internal Medicine, Nishinomiya Kaisei Hospital Hyogo, Japan
| |
Collapse
|
6
|
Wakabayashi I, Masuda H. Association of pulse pressure with carotid atherosclerosis in patients with type 2 diabetes mellitus. Blood Press 2007; 16:56-62. [PMID: 17453753 DOI: 10.1080/08037050701248051] [Citation(s) in RCA: 5] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/23/2022]
Abstract
Pulse pressure is known to be a potent risk factor for cardiovascular disease. However, it has not been determined whether pulse pressure is associated with cerebrovascular disease in patients with diabetes mellitus. The purpose of the present study was to investigate association of pulse pressure with carotid atherosclerosis in patients with diabetes and to determine whether age and gender affect the association. A cross-sectional study was performed using outpatients with type 2 diabetes. Carotid atherosclerosis and arterial stiffness were evaluated by intima-media thickness of common carotid arteries (IMT) and aortic pulse wave velocity (a-PWV), respectively. The relationship of brachial pulse pressure with IMT was analyzed. Pulse pressure was significantly correlated with age, duration of diabetes, systolic blood pressure, serum sialic acid, a-PWV and IMT. The mean IMT in the highest tertile group of pulse pressure after adjustment for gender was significantly higher than the mean IMT in the lowest tertile group. In analysis using subgroups of gender, a significant correlation between pulse pressure and IMT was shown in women but not in men. In multiple regression analysis, there was a significant correlation between pulse pressure and IMT in women independently of age, body mass index, systolic blood pressure, hemoglobin A1c, atherogenic index, a-PWV, history of therapy for hypertension and history of nephropathy. The results of this study suggest that pulse pressure is a major age-independent predictor of carotid atherosclerosis in female patients with diabetes.
Collapse
Affiliation(s)
- Ichiro Wakabayashi
- Department of Environmental and Preventive Medicine, Hyogo College of Medicine, Mukogawa-cho 1-1, Nishinomiya, Hyogo 663-8501, Japan.
| | | |
Collapse
|
7
|
Wakabayashi I, Masuda H. Association of pulse pressure with fibrinolysis in patients with type 2 diabetes. Thromb Res 2007; 121:95-102. [PMID: 17439823 DOI: 10.1016/j.thromres.2007.02.014] [Citation(s) in RCA: 12] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/09/2007] [Revised: 02/20/2007] [Accepted: 02/27/2007] [Indexed: 11/15/2022]
Abstract
INTRODUCTION Pulse pressure is a potent risk factor for atherosclerotic disease. The purpose of this cross-sectional study was to determine whether pulse pressure is associated with blood coagulation and fibrinolysis in patients with diabetes. MATERIALS AND METHODS The relationships between pulse pressure and atherosclerotic risk factors, including blood coagulation and fibrinolysis markers, were investigated in subjects with type 2 diabetes. RESULTS Pulse pressure was found to be significantly correlated with blood fibrinolysis markers, such as plasmin-alpha2-antiplasmin complex (PAP) and D-dimer, but not with platelets and blood coagulation markers, such as fibrinogen and thrombin-antithrombin III complex (TAT). The mean pulse pressure levels in the highest tertiles of PAP and D-dimer were significantly higher than those in the lowest tertiles, while the differences in the mean pulse pressure levels among tertile groups of platelets, fibrinogen and TAT were not significant. These relationships of pulse pressure with PAP and D-dimer were not altered by adjustment for age, sex and history of therapy with antihypertensive and/or antithrombotic drugs. On the other hand, neither systolic nor diastolic blood pressure showed significant correlations with PAP, D-dimer, platelets, fibrinogen and TAT. Pulse pressure was also significantly correlated with age, aortic pulse wave velocity, intima-media thickness of the common carotid arteries and serum sialic acid. CONCLUSIONS Pulse pressure is associated with blood fibrinolysis as well as atherosclerotic progression in patients with type 2 diabetes. The results of this study suggest that pulse pressure affects atherosclerotic progression through altering hemostatic functions in patients with diabetes.
Collapse
Affiliation(s)
- Ichiro Wakabayashi
- Department of Environmental and Preventive Medicine, Hyogo College of Medicine, Mukogawa-cho 1-1, Nishinomiya, Hyogo 663-8501, Japan.
| | | |
Collapse
|
8
|
|
9
|
Rajala U, Laakso M, Päivänsalo M, Pelkonen O, Ruokonen A, Suramo I, Keinänen-Kiukaanniemi S. Associations of microalbuminuria and blood pressure with carotid, aortic and femoral atheromatous plaques in elderly Finns. Diabetes Res Clin Pract 2005; 69:262-71. [PMID: 16098923 DOI: 10.1016/j.diabres.2005.01.001] [Citation(s) in RCA: 5] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 11/24/2004] [Revised: 12/20/2004] [Accepted: 01/11/2005] [Indexed: 11/22/2022]
Abstract
AIMS To evaluate the possible associations of microalbuminuria (MA) and blood pressure (BP) with the ultrasonographic manifestations of carotid, aortic and femoral atherosclerosis in 65-year-old Finns. METHODS Ultrasonographic measurements were performed on 54 diabetic subjects, 97 subjects with impaired glucose tolerance (IGT) and 57 normoglycemic subjects (NGT). Urinary albumin and creatinine concentrations were measured from an early morning spot urine sample, and the urinary albumin-to-creatinine ratio (ACR) of > or = 2.5 mg/mmol in men and > or = 3.5 mg/mmol in women was used as a measure of MA. Hypertension was defined as either a systolic BP of > or = 160 mmHg or a diastolic BP of > or = 95 mmHg or being on antihypertensive medication. RESULTS Eighteen subjects were microalbuminuric and 176 subjects normoalbuminuric. MA was associated with diabetes mellitus and high systolic and diastolic BP. The subjects were divided into two groups according to the median total number of carotid, aortic and femoral plaques: > or = 9 versus 0-8 plaques. A high number of plaques were associated with hypertension, male gender, smoking and MA. When the study subjects were stratified according to hypertension, it turned out that MA was associated with a high number of plaques in hypertensive, but not in nonhypertensive subjects. According to the results of logistic regression analysis with a high number of plaques as the dependent variable, the unadjusted OR for smoking was 6.0 (95% CI 2.4-15.3) in hypertensive subjects. Microalbuminuria was of borderline statistical significance (OR 4.5, 95% CI 0.9-22.9). After adjustment for systolic blood pressure and fasting glucose concentration, the OR for microalbuminuria was reduced to 3.3 (95% CI 0.6-18.4).
Collapse
Affiliation(s)
- Ulla Rajala
- Department of Public Health Science and General Practice, University of Oulu, Aapistie 1, 90220 Oulu, Finland.
| | | | | | | | | | | | | |
Collapse
|
10
|
Rajala U, Laakso M, Päivänsalo M, Suramo I, Keinänen-Kiukaanniemi S. Blood pressure and atherosclerotic plaques in carotid, aortic and femoral arteries in elderly Finns with diabetes mellitus or impaired glucose tolerance. J Hum Hypertens 2004; 19:85-91. [PMID: 15343356 DOI: 10.1038/sj.jhh.1001779] [Citation(s) in RCA: 12] [Impact Index Per Article: 0.6] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/09/2022]
Abstract
The aim of the study was to evaluate the occurrence of atheromatous plaques in carotid, aortic and femoral arteries, focusing on blood pressure (BP). The study subjects consisted of 65-year-old Finns drawn from a population-based cohort. Ultrasonographic measurements were performed on 54 diabetic subjects, 97 subjects with impaired glucose tolerance (IGT) and 57 normoglycaemic subjects (NGT). High systolic BP (SBP) was defined as >or=160 mmHg and high diastolic BP as >or=95 mmHg. High pulse pressure (PP) was defined as the highest tertile (>or=75 mmHg) of PP and high mean BP (MBP) as the highest tertile (>or=111 mmHg) of MBP. The prevalence of atheromatous plaques was 77% (160/208) in carotid arteries, 94% (195/208) in aorta and 77% (161/208) in femoral arteries. A total of 64% (134/208) of the subjects had plaques in both carotid and femoral arteries, and they were compared with those who had plaques in 0-1 of these arteries. In addition to male gender and long-lasting smoking, the occurrence of plaques in both carotid and femoral arteries were associated with high SBP and high MBP. According to the results of multiple regression analyses, the adjusted odds ratio for plaques in both carotid and femoral arteries was 3.1 (95% CI 1.5-6.5) in subjects with high SBP compared to those with lower SBP. When SBP was replaced by high MBP, the adjusted odds ratio for it was 2.3 (95% CI 1.1-4.8).
Collapse
Affiliation(s)
- U Rajala
- Department of Public Health Science and General Practice, University of Oulu, Oulu, Finland.
| | | | | | | | | |
Collapse
|