Search

ABSTRACT Introduction The early detection of high‐risk individuals is crucial to delay and reduce the incidence of type 2 diabetes. In this study, we aimed to explore the performance of a novel subgroup‐specific biomarker strategy in the prediction of incident diabetes. Materials and Methods In the Taiwan Lifestyle Cohort Study, adult subjects without diabetes were included and followed for the incidence of diabetes in 2006–2019. The biomarkers measured included blood secretogranin III (SCG3), vascular adhesion protein‐1 (VAP‐1), fibrinogen‐like protein 1 (FGL1), angiopoietin‐like protein 6 (ANGPTL6), and angiopoietin‐like protein 4 (ANGPTL4). Results Among the 1,287 subjects, 12.2% developed diabetes during a 6 year follow‐up. Blood VAP‐1 was significantly associated with incident diabetes in the overall population (HR = 0.724, P

Abstract The prevalence of osteoporosis has drastically increased recently. It is not only the most frequent but is also a major global public health problem due to its high morbidity. There are many risk factors associated with osteoporosis were identified. However, most studies have used the traditional multiple linear regression (MLR) to explore their relationships. Recently, machine learning (Mach-L) has become a new modality for data analysis because it enables machine to learn from past data or experiences without being explicitly programmed and could capture nonlinear relationships better. These methods have the potential to outperform conventional MLR in disease prediction. In the present study, we enrolled a Chinese post-menopause cohort followed up for 4 years. The difference of T-score (δ-T score) was the dependent variable. Information such as demographic, biochemistry and life styles were the independent variables. Our goals were: (1) Compare the prediction accuracy between Mach-L and traditional MLR for δ-T score. (2) Rank the importance of risk factors (independent variables) for prediction of δ T-score. Totally, there were 1698 postmenopausal women were enrolled from MJ Health Database. Four different Mach-L methods namely, Random forest (RF), eXtreme Gradient Boosting (XGBoost), Naïve Bayes (NB), and stochastic gradient boosting (SGB), to construct predictive models for predicting δ-BMD after four years follow-up. The dataset was then randomly divided into an 80% training dataset for model building and a 20% testing dataset for model testing. A 10-fold cross-validation technique for hyperparameter tuning was used. The model with the lowest root mean square error for the validation dataset was viewed as the best model for each ML method. The averaged metrics of the RF, SGB, NB, and XGBoost models were used to compare the model performance of the benchmark MLR model that used the same training and testing dataset as the Mach-L methods. We defined that the priority demonstrated in each model ranked 1 as the most critical risk factor and 22 as the last selected risk factor. For Pearson correlation, age, education, BMI, HDL-C, and TSH were positively and plasma calcium level, and baseline T-score were negatively correlated with δ-T score. All four Mach-L methods yielded lower prediction errors than the MLR method and were all convincing Mach-L models. From our results, it could be noted that education level is the most important factor for δ-T Score, followed by DBP, smoking, SBP, UA, age, and LDL-C. All four Mach-L outperformed traditional MLR. By using Mach-L, the most important six risk factors were selected which are, from the most important to the least: DBP, SBP, UA, education level, TG and sleeping hour. δ T score was positively related to SBP, education level, UA and TG and negatively related to DBP and sleeping hour in postmenopausal Chinese women.

ObjectivePregnancies with large-for-gestational-age (LGA) fetuses are associated with increased risks of various adverse perinatal outcomes. While existing research primarily focuses on term neonates, less is known about preterm neonates. This study aims to explore the risks of adverse maternal and neonatal perinatal outcomes associated with LGA in term neonates and neonates with different degrees of prematurity, compared to appropriate-for-gestational-age (AGA) neonates.MethodsUsing the Birth Reporting Databases (2007–2018) linked to Taiwan's National Health Insurance Research Database, we conducted a retrospective nationwide cohort study of singleton neonates delivered between 24 and 42 weeks of gestation. Based on gestational age at delivery, the enrolled neonates were classified into term (37–42 weeks of gestation), late preterm (34–36 weeks of gestation), moderate preterm (32–33 weeks of gestation), very preterm (28–31 weeks of gestation), and extremely preterm (24–27 weeks of gestation). LGA was defined by the 2013 World Health Organization (WHO) growth standard and the Taiwan growth standard. Perinatal outcomes were compared between LGA and AGA neonates across different gestational age groups.ResultsAmong the 1,602,638 neonates, 44,359 were classified as LGA by the 2013 WHO growth standard. Compared to AGA neonates, LGA neonates in term and late preterm groups exhibited higher risks of primary cesarean section, prolonged labor, neonatal hypoglycemia, birth trauma, hypoxic ischemic encephalopathy, jaundice needing phototherapy, respiratory distress, neonatal intensive care unit (NICU) admission, newborn sepsis, and fetal death. However, most of these risks were not increased in moderate, very, and extremely preterm groups. Conversely, being LGA was associated with lower risks of primary cesarean section (very preterm group), jaundice needing phototherapy (moderate and very preterm groups), respiratory distress (moderate and very preterm groups), NICU admission (moderate and very preterm groups), newborn sepsis (very preterm group), retinopathy of prematurity (late, moderate, and very preterm groups), and bronchopulmonary dysplasia (very preterm group). These findings remained consistent when the Taiwan growth standard was applied.ConclusionBeing LGA is associated with increased risks of perinatal complications in term and late preterm neonates, but not in earlier preterm groups. These findings underscore the importance of tailoring management strategies for LGA neonates to consider different degrees of prematurity.

Abstract Introduction The prevalence of type 2 diabetes (T2D) has increased dramatically in recent decades, and there are increasing indications that dementia is related to T2D. Previous attempts to analyze such relationships principally relied on traditional multiple linear regression (MLR). However, recently developed machine learning methods (Mach-L) outperform MLR in capturing non-linear relationships. The present study applied four different Mach-L methods to analyze the relationships between risk factors and cognitive function in older T2D patients, seeking to compare the accuracy between MLR and Mach-L in predicting cognitive function and to rank the importance of risks factors for impaired cognitive function in T2D. Methods We recruited older T2D between 60–95 years old without other major comorbidities. Demographic factors and biochemistry data were used as independent variables and cognitive function assessment (CFA) was conducted using the Montreal Cognitive Assessment as an independent variable. In addition to traditional MLR, we applied random forest (RF), stochastic gradient boosting (SGB), Naïve Byer’s classifier (NB) and eXtreme gradient boosting (XGBoost). Results Totally, the test cohort consisted of 197 T2D (98 men and 99 women). Results showed that all ML methods outperformed MLR, with symmetric mean absolute percentage errors for MLR, RF, SGB, NB and XGBoost respectively of 0.61, 0.599, 0.606, 0.599 and 0.2139. Education level, age, frailty score, fasting plasma glucose and body mass index were identified as key factors in descending order of importance. Conclusion In conclusion, our study demonstrated that RF, SGB, NB and XGBoost are more accurate than MLR for predicting CFA score, and identify education level, age, frailty score, fasting plasma glucose, body fat and body mass index as important risk factors in an older Chinese T2D cohort.

BackgroundVascular adhesion protein-1 (VAP-1), a dual-function glycoprotein, has been reported to play a crucial role in inflammation and tumor progression. We conducted a community-based cohort study to investigate whether serum VAP-1 could be a potential biomarker for predicting incident cancers and mortality.MethodFrom 2006 to 2018, we enrolled 889 cancer-free subjects at baseline. Serum VAP-1 levels were measured using a time-resolved immunofluorometric assay. Cancer and vital status of the participants were obtained by linking records with the computerized cancer registry and death certificates in Taiwan.ResultsDuring a median follow-up of 11.94 years, 69 subjects developed incident cancers and 66 subjects died, including 29 subjects who died from malignancy. Subjects in the highest tertile of serum VAP-1 had a significantly higher risk of cancer incidence (p=0.0006), cancer mortality (p=0.0001), and all-cause mortality (p=0.0002) than subjects in the other tertiles. The adjusted hazard ratios per one standard deviation increase in serum VAP-1 concentrations were 1.28 for cancer incidence (95% CI=1.01–1.62), 1.60 for cancer mortality (95% CI=1.14–2.23), and 1.38 for all-cause mortality (95% CI=1.09–1.75). The predictive performance of serum VAP-1 was better than that of gender, smoking, body mass index, hypertension, diabetes, and estimated glomerular filtration rate but lower than that of age for cancer incidence, cancer mortality, and all-cause mortality, as evidenced by higher increments in concordance statistics and area under the receiver operating characteristic curve.ConclusionSerum VAP-1 levels are associated with a 12-year risk of incident cancer, cancer mortality, and all-cause mortality in a general population.

Aim: Several studies have demonstrated that factors including diabetes, including insulin resistance (IR), glucose effectiveness (GE), and the first and second phase of insulin secretion (FPIS, SPIS) could easily be calculated using basic characteristics and biochemistry profiles. Aging is accompanied by deteriorations of insulin resistance (IR) and insulin secretion. However, little is known about the roles of aging in the different phases of insulin secretion (ISEC), i.e., the first and second phase of insulin secretion (FPIS, SPIS), and glucose effectiveness (GE). Methods: In total, 169 individuals (43 men and 126 women) recruited from the data bank of the Meei-Jaw (MJ) Health Screening Center and Cardinal Tien Hospital Data Access Center between 1999 and 2008, with a similar fasting plasma glucose (FPG: 90 mg/dL) and BMI (men: 23 kg/m2, women 22 kg/m2) were enrolled. The IR, FPIS, SPIS, and GE were estimated using our previously developed equations shown below. Pearson correlation analysis was conducted to assess the correlations between age and four diabetes factors (DFs: IR, FPIS, SPIS, and GE). The equations that are used to calculate the DF in the present study were built and published by our group. Results: The age of the participants ranged from 18 to 78 years. Men had higher FPIS but lower HDL-C levels than women (2.067 ± 0.159, 1.950 ± 0.186 μU/min and 1.130 ± 0.306, 1.348 ± 0.357 mmol/dl, accordingly). The results of the Pearson correlation revealed that age was negatively related to the IR and GE in both genders (IR: r = −0.39, p < 0.001 for men, r = −0.24, p < 0.003 for women; GE: r = 0.66, p < 0.001 for men, r = 0.78, p < 0.001 for women). At the same time, the FPIS was also only found to be negatively correlated with age in females (r = −0.238, p = 0.003), but there was no difference in the SPIS and age among both genders. Conclusions: We have found that in Chinese subjects with a normal FPG level (90 mg/dL) and body mass index (men: 23 kg/m2, women: 22: kg/m2), age is negatively related to the IR and GE among both genders. Only the FPIS was found to be negatively related to age in women. The tightness of their relationships, from the highest to the lowest, are GE, FPIS, and IR. These results should be interpreted with caution because of the small sample size.

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.

Type 2 diabetes mellitus (T2DM) patients have a high risk of coronary artery disease (CAD). Thallium-201 myocardial perfusion scan (Th-201 scan) is a non-invasive and extensively used tool in recognizing CAD in clinical settings. In this study, we attempted to compare the predictive accuracy of evaluating abnormal Th-201 scans using traditional multiple linear regression (MLR) with four machine learning (ML) methods. From the study, we can determine whether ML surpasses traditional MLR and rank the clinical variables and compare them with previous reports.In total, 796 T2DM, including 368 men and 528 women, were enrolled. In addition to traditional MLR, classification and regression tree (CART), random forest (RF), stochastic gradient boosting (SGB) and eXtreme gradient boosting (XGBoost) were also used to analyze abnormal Th-201 scans. Stress sum score was used as the endpoint (dependent variable). Our findings show that all four root mean square errors of ML are smaller than with MLR, which implies that ML is more precise than MLR in determining abnormal Th-201 scans by using clinical parameters. The first seven factors, from the most important to the least are:body mass index, hemoglobin, age, glycated hemoglobin, Creatinine, systolic and diastolic blood pressure. In conclusion, ML is not inferior to traditional MLR in predicting abnormal Th-201 scans, and the most important factors are body mass index, hemoglobin, age, glycated hemoglobin, creatinine, systolic and diastolic blood pressure. ML methods are superior in these kinds of studies.

Cushing's disease (CD) is the most common cause of endogenous Cushing's syndrome. Transsphenoidal surgery (TSS) is the first choice of treatment. Predicting prognosis after treatment can benefit further strategies of management, but currently there is no convenient predictor. This study aims to investigate characteristic changes after treatment and to identify potential prognostic predictors. Methods: We retrospectively studied the records of CD patients presenting to the National Taiwan University Hospital, Taipei, Taiwan between 1992 and 2011. They were categorized according to treatment response. Clinical features and examination findings were compared between groups. Results: Forty-one patients with CD were included. The follow-up time was 0.26–19.3 years. The time interval between the onset of symptoms and diagnosis was 2.1–120.0 months. The initial remission rate of CD after the first treatment was 82.9%. Mean body mass index (BMI) was 27.4 kg/m2 before treatment and 26.0 kg/m2 3 months after treatment. The patients in remission had a greater decrease in BMI after treatment and lower dehydroepiandrosterone sulfate (DHEAS) levels before treatment, compared with the recurrent group (both p

AimThe addition of maternal age to fasting plasma glucose (FPG) at 24-28 gestational weeks improves the performance of GDM screening as maternal age increases. However, this method delays the diagnosis of GDM. Since FPG at the first prenatal visit (FPV) is a screening option for pre-existing diabetes, we evaluated the performance of age plus FPG at the FPV to reduce the need for the OGTT.MethodsPregnant women were recruited consecutively in 2013-2018 (the training cohort) and 2019 (the validation cohort). We excluded women with twin pregnancies, unavailable FPG at the FPV or OGTT data, pre-pregnancy diabetes, or a history of GDM. All participants underwent FPG and haemoglobin A1c (HbA1c) at the FPV and received 75-g OGTT at 24-28 gestational weeks if FPG at the FPV was ResultsThe incidence of GDM increased with age. The "FPG at the FPV" algorithm reduced OGTT% to 68.8% with the FPG cutoff at 79 mg/dl. The "age plus FPG at the FPV" algorithm, with the cutoff of 114, further reduced OGTT% to 58.3%, with the sensitivity of 90.7% (9.3% GDM missed) and the specificity of 100%. These findings were replicated in the validation cohort.ConclusionsScreening GDM by maternal age plus FPG at the FPV can reduce OGTT%, especially in populations with a significant proportion of pregnant women with advanced ages.

AimOverweight and obesity are important risk factors of gestational diabetes mellitus (GDM). Clustering of metabolic risk factors in early pregnancy may be a potential pathogenesis between the link of overweight/obesity and GDM. Since it remains unexplored, we investigated if overweight and obesity are associated with clustering of metabolic risk factors in early pregnancy and the risk of GDM in this cohort study.MethodsTotal 527 women who visited National Taiwan University Hospital for prenatal care in between November 2013 to April 2018 were enrolled. Risk factors of GDM in the first prenatal visit (FPV) were recorded. Overweight/obesity was defined if body mass index ≥24 kg/m2. GDM was diagnosed from the result of a 75g oral glucose tolerance test in 24-28 gestational weeks.ResultsOverweight/obesity was associated with clustering of metabolic risk factors of GDM, including high fasting plasma glucose, high HbA1c, insulin resistance, high plasma triglyceride and elevated blood pressure in FPV (pConclusionsOverweight/obesity is associated with clustering of metabolic risk factors in early pregnancy, which is correlated with higher risk of GDM. Our findings suggest that metabolic risk factors during early pregnancy should be evaluated in overweight/obese women.

OBJECTIVE:Using a specific cutoff of fasting plasma glucose (FPG) to screen gestational diabetes mellitus (GDM) can reduce the use of oral glucose tolerance tests (OGTT). Since the prevalence of GDM increases with age, this screening method may not be appropriate in healthcare systems where women become pregnant at older ages. Therefore, we aimed to develop a screening algorithm for GDM that takes maternal age into consideration. METHODS:We included 945 pregnant women without history of GDM who received 75g OGTT to diagnose GDM in 2011. Screening algorithms using FPG with or without age were developed. Another 362 pregnant women were recruited in 2013-2015 as the validation cohort. RESULTS:Using FPG criteria alone, more GDM diagnoses were missed in women ≥35 years than in women

Context: Maternal lipids during pregnancy and placental growth factors are associated with excess fetal growth. However, how these factors interact to increase the risk of delivering large-for-gestational-age (LGA) neonates remains unclear. In this study, we investigated the relationship between maternal plasma triglycerides (TGs) and free fatty acids (FFAs) during pregnancy, cord blood insulin-like growth factors (IGF), and LGA.Objective: In a cell model, we studied the effect of different FAs on placental IGF-1 secretion.Methods: This cohort study included pregnant women with term pregnancy and without diabetes or hypertensive disorders in pregnancy. Maternal fasting plasma TGs and FFAs were measured in the second trimester. Cord blood IGF-1, IGF-2, and IGF binding protein-1 and protein-3 were measured at the time of delivery. A human trophoblast cell line, 3A-sub-E, was used to evaluate the effect of different FFAs on placental IGF-1 secretion.Results: We recruited 598 pregnant women-neonate pairs. Maternal plasma TG (180 mg/dL [152.5-185.5 mg/dL] vs 166 mg/dL [133-206 mg/dL], P = .04) and cord blood IGF-1 concentrations (72.7 ± 23.0 vs 54.1 ± 22.8 ng/mL, P < .001) were higher in the LGA group and were significantly associated with birth weight z score. Maternal plasma free palmitic acid (PA) and stearic acid (SA), but not oleic acid (OA) or linoleic acid (LA), were significantly associated with cord blood IGF-1 concentrations. In 3A-sub-E cells, treatment with PA, SA, and LA, but not OA, induced IGF-1 expression and secretion.Conclusion: Certain FFAs can induce placental IGF-1 secretion, which suggests a potential pathophysiology linking maternal plasma lipids and LGA. [ABSTRACT FROM AUTHOR]