Association of sodium-glucose cotransporter 2 inhibitors with risk of incident dementia and all-cause mortality in older patients with type 2 diabetes: A retrospective cohort study using the TriNetX US collaborative networks.

BACKGROUND: Limited evidence exists to support any specific medication over others to prevent dementia in older patients with type 2 diabetes (T2D). We investigated whether treatment with sodium-glucose cotransporter 2 (SGLT-2) inhibitors is associated with a lower risk of incident dementia and all-cause mortality, relative to dipeptidyl peptidase-4 (DPP-4) inhibitors and glucagon-like peptide-1 receptor agonists (GLP-1 RA). METHODS: In this retrospective, active-comparator cohort study, we used data from the TriNetX electronic health records network. Our primary cohort comprised patients with T2D aged ≥50 years, registered between January 2012 and December 2022. Patients with a history of dementia were excluded. We used Kaplan-Meier survival analysis to estimate the incidence of dementia and all-cause mortality in our cohort after they had used glucose-lowering drugs for at least 12 months. Propensity score matching was performed to balance the SGLT-2 inhibitor, DPP-4 inhibitor and GLP-1 RA cohorts. Subgroup analyses for sex and age were also conducted. RESULTS: Our first cohort comprised 193 948 patients treated with metformin and SGLT-2 inhibitors and an equal number of patients treated with metformin and DPP-4 inhibitors. In this cohort, the risk of dementia and all-cause mortality was lower in patients treated with SGLT-2 inhibitors than in those treated with DPP-4 inhibitors (hazard ratio [HR]: 0.62, 95% confidence interval [CI]: 0.59-0.65, for dementia; HR: 0.54, 95% CI: 0.52-0.56, for all-cause mortality). Our second cohort comprised 165 566 patients treated with metformin and SGLT-2 inhibitors and an equal number of patients treated with metformin and GLP-1 RAs. In this cohort, the risk of dementia and all-cause mortality was lower in those treated with SGLT-2 inhibitors than in those treated with GLP-1 RAs (HR: 0.92, 95% CI: 0.87-0.98, for dementia; HR: 0.88, 95% CI: 0.85-0.91, for all-cause mortality). CONCLUSIONS: The use of SGLT-2 inhibitor was associated with a lower risk of incident dementia and all-cause mortality in older adults with T2D compared to DPP-4 inhibitor and GLP-1 RA.

Machine Learning Models for Predicting Mortality in Patients with Cirrhosis and Acute Upper Gastrointestinal Bleeding at an Emergency Department: A Retrospective Cohort Study.

BACKGROUND: Cirrhosis is a major global cause of mortality, and upper gastrointestinal (GI) bleeding significantly increases the mortality risk in these patients. Although scoring systems such as the Child-Pugh score and the Model for End-stage Liver Disease evaluate the severity of cirrhosis, none of these systems specifically target the risk of mortality in patients with upper GI bleeding. In this study, we constructed machine learning (ML) models for predicting mortality in patients with cirrhosis and upper GI bleeding, particularly in emergency settings, to achieve early intervention and improve outcomes. METHODS: In this retrospective study, we analyzed the electronic health records of adult patients with cirrhosis who presented at an emergency department (ED) with GI bleeding between 2001 and 2019. Data were divided into training and testing sets at a ratio of 90:10. The ability of three ML models-a linear regression model, an XGBoost (XGB) model, and a three-layer neural network model-to predict mortality in the patients was evaluated. RESULTS: A total of 16,025 patients with cirrhosis and 32,826 ED visits for upper GI bleeding were included in the study. The in-hospital and ED mortality rates were 11.2% and 2.2%, respectively. The XGB model exhibited the highest performance in predicting both in-hospital and ED mortality (area under the receiver operating characteristic curve: 0.866 and 0.861, respectively). International normalized ratio, renal function, red blood cell distribution width, age, and white blood cell count were the strongest predictors in all the ML models. The median ED length of stay for the ED mortality group was 17.54 h (7.16-40.01 h). CONCLUSIONS: ML models can be used to predict mortality in patients with cirrhosis and upper GI bleeding. Of the three models, the XGB model exhibits the highest performance. Further research is required to determine the actual efficacy of our ML models in clinical settings.

Glycemic variability's impact on painful diabetic peripheral neuropathy in type 2 diabetes patients.

Hyperglycemia in type 2 diabetes leads to diabetic peripheral neuropathy (DPN) and neuropathic pain, yet the association between glycemic variability and painful DPN remains insufficiently evidenced. To address this, we conducted a prospective longitudinal cohort study involving adult type 2 diabetes patients at a medical center. DPN was identified using the Michigan Neuropathy Screening Instrument (MNSI), and neuropathic pain was assessed with the Taiwan version of the Douleur Neuropathique 4 (DN4-T) questionnaire. At baseline in 2013, all participants were free of DPN and were re-evaluated in 2019 for the development of painful DPN. We measured visit-to-visit glycemic fluctuations using the coefficient of variation (CV) of fasting plasma glucose (FPG) and glycated hemoglobin (HbA1c). Patients were stratified into tertiles according to their FPG-CV and HbA1c-CV. Among the 622 participants, 267 developed DPN during the six-year follow-up. Following matching of age and sex, 210 patients without DPN and 210 with DPN (including 26 with neuropathic pain) were identified. Our findings revealed a significant association between high FPG-CV and painful DPN, with the highest tertile showing an adjusted odds ratio of 2.82 (95% confidence interval 1.04-7.64) compared to the lowest tertile. On the contrary, HbA1c-CV did not show a significant association with the risk of painful DPN. Our study indicates that higher FPG-CV is associated with an increased risk of painful DPN, supporting the role of glycemic variability in the development of painful DPN.

Maternal autoimmune disease and offspring risk of haematological malignancies: a case-control study.

Background: Autoimmune diseases are known to be associated with an increased risk of cancer. Whether maternal immune dysregulation can have an impact on the development of haematological malignancies in offspring remains uncertain. Therefore, we explored the association between offspring risk of haematological malignancies and maternal autoimmune disease using a real-world nationwide population-based study. Methods: In this case-control study, we identified 2172 children with haematological malignancies between 2004 and 2019 from Taiwan's National Health Insurance program and compared them with population-based controls without haematologic malignancies, who were matched with each individual at a ratio of 1:4. The medical information of the autoimmune mothers were obtained from the Taiwan Maternal and Child Health Database. Conditional logistic regression was used to estimate the odds ratio for haematologic malignancy in offspring. Furthermore, subgroup and stratified analyses were conducted. Findings: Among the rheumatologic diseases in our study, Crohn's disease was the most common disease both in the haematological malignancy group (1.1%) and the control group (0.9%). In multivariable analysis, the odds ratio for haematological malignancy in offspring with maternal autoimmune diseases was 1.2 (95% confidence interval [CI] 0.91-1.58). The overall risk of haematologic malignancy was not significantly higher when adjusted for specific risk factors, including neonatal age, maternal age, family income, urbanization, maternal occupation, birth weight, or maternal comorbidity, except for prematurity. When comparing different autoimmune diseases among haematological malignancies and the control group, maternal psoriatic arthritis/psoriasis had the highest adjusted overall risk for haematological malignancies (adjusted OR 2.11, CI 0.89-5), followed by ankylosing spondylitis (adjusted OR 1.45, CI 0.7-3), autoimmune thyroiditis (OR 1.26, CI 0.57-2.81), systemic lupus erythematosus (OR 1.21, CI 0.48-3.02), Crohn's disease (OR 1.19, CI 0.75-1.9), and Sjogren's syndrome (OR 1.18, CI 0.65-2.15), but no significance was reached in these analyses. Multivariable analysis of risk factors associated with haematological malignancy subtypes was done. It showed no associations between maternal autoimmune disease and childhood haematological malignancies. Interpretation: We found no significant relationship between maternal autoimmune disease and childhood haematological malignancies. The influence of maternal immune dysregulation on the next generation with respect to haematological malignancies development may be limited. Funding: There was no funding source for this study.

Association of the ABCG2 rs2231142 variant with the Framingham Cardiovascular Disease Risk score in the Taiwanese population.

Background: Serum uric acid (SUA) is an important predictor of cardiovascular events and mortality. The ABCG2 rs2231142 variant (TT genotype) is associated with hyperuricemia (HUA), but the relationship between ABCG2 gene polymorphisms and coronary artery disease (CAD) risk is poorly elucidated. We investigated the association between ABCG2 rs2231142 genetic variants and the Framingham Risk Score for Cardiovascular Disease (FRS-CVD) in a Taiwanese population. Methods: This cross-sectional study enrolled 139,508 Taiwanese participants aged 30-70 years based on data from the Taiwan Biobank (TWB) database that was obtained from questionnaires, laboratory investigations, anthropometry, and Affymetrix TWB genome-wide single-nucleotide polymorphism (SNP) chip data analysis. The association between ABCG2 rs2231142 and FRS-CVD risk was evaluated using logistic regression analysis. Results: Compared to those with the GG genotype, participants with the ABCG2 rs2231142 TT genotype had a significantly lower systolic blood pressure, smoking rate, body mass index, triglyceride level, waist circumference, waist-hip ratio, and body fat percentage, but had higher high-density lipoprotein cholesterol level. Despite the same FRS-CVD score, participants with TT genotypes had higher SUA. Even with the same SUA, TT carriers had a lower FRS-CVD than GT and GG carriers. Participants with the TT genotype had significantly lower CVD risk, particularly female participants with HUA and BMI <27 (OR: 0.760, 95 % CI: 0.587-0.985; p = 0.0381) group. Conclusion: The ABCG2 rs2231142 TT genotype is associated with a lower FRS-CVD, particularly in non-obese hyperuricemic female individuals. The complicated interplay among genetic variations, metabolic profile, and CVD risk provides insights for precision health.

Influenza Infection During Pregnancy and Risk of Seizures in Offspring.

Importance: Seizure is a common neurological problem among infants and children up to age 6 years. Prenatal exposure to maternal influenza infection has been reported to be associated with childhood seizures. Objective: To evaluate the association between maternal influenza infection and risk of childhood seizures. Designs, Setting, and Participants: This cohort study identified mother-offspring pairs from January 1, 2004, to December 31, 2013, using records in Taiwan's Maternal and Child Health Database. Mothers who had influenza infection during pregnancy and their first offspring were identified and assigned to the influenza group. Mothers in the control group were those without influenza during pregnancy and were matched 1:4 with mothers in the influenza group by maternal age, offspring sex, and date of delivery. Offspring were followed up until December 31, 2020. Data were analyzed between March 2023 and July 2024. Exposure: Diagnosis of influenza infection during pregnancy defined using International Classification of Diseases, Ninth Revision, Clinical Modification codes 487.0, 487.1, and 487.8, or International Statistical Classification of Diseases, Tenth Revision, Clinical Modification codes J09, J10, and J11. Main Outcomes and Measures: The primary outcome was the association between maternal influenza infection during pregnancy and risk of any type of seizures during childhood, including both epilepsy and febrile seizures. Adjusted hazard ratios (AHRs) and 95% CIs were estimated using Cox proportional hazards regression models. Pregnancy-related complications were collected as covariates. Results: A total of 1 316 107 mother-offspring pairs were enrolled, of whom 75 835 mothers (predominant maternal age, 25-29 years; 39 324 male offspring [51.9%]) were assigned to the influenza group and 1 240 272 were matched and assigned to the control group (n = 303 340; predominant maternal age, 30-34 years; 157 296 male offspring [51.9%]). In the influenza group, there was a slightly higher prevalence of placenta previa or abruption compared with the control group (1.6% [1241] vs 1.4% [4350]; P < .001). The cumulative risk of seizures was higher among offspring whose mothers had influenza infection. After controlling for potential confounders, the AHRs were 1.09 (95% CI, 1.05-1.14) for seizures, 1.11 (95% CI, 1.06-1.17) for febrile convulsions, and 1.04 (95% CI, 0.97-1.13) for epilepsy. In the subgroup analysis, no statistically significant differences were observed between the trimesters regarding the timing of influenza infection. Conclusions and Relevance: Results of this cohort study suggest that maternal influenza infection during pregnancy was associated with an increased risk of childhood seizures, especially febrile seizures, but not epilepsy. Further studies are needed to elucidate the mechanisms underlying childhood neurological development.

Use of Angiotensin-Converting Enzyme Inhibitors and Angiotensin Receptor Blockers Is Associated With a Reduced Risk of Poststroke Epilepsy in Patients With Ischemic Stroke.

BACKGROUND: Poststroke epilepsy (PSE) is a common complication after ischemic stroke. This study investigates the association between the use of angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs) and the risk of PSE in patients with ischemic stroke. METHODS AND RESULTS: A population-based retrospective cohort study was conducted using Taiwan's National Health Insurance Research Database between 2000 and 2015. Patients with hypertension with a history of ischemic stroke were classified into prevalent, new, and nonusers according to their use of ACEIs/ARBs. Prevalent ACEI/ARB users were further classified into continuing or discontinued users, based on their poststroke medication adherence. We used multivariate Cox regression models, adjusted for demographics and comorbidities, to assess the risk of PSE among different ACEI/ARB user groups. There were 182 983 ACEI/ARB users and 38 365 nonusers included. There were 7387 patients diagnosed with PSE, whereas 213 961 were not. Nonusers exhibited a higher risk of PSE (adjusted hazard ratio [aHR], 1.72 [95% CI, 1.63-1.82]). Both prevalent and nonusers had higher risks compared with new ACEI/ARB users, with respective aHRs of 1.33 (95% CI, 1.25-1.41) and 2.00 (95% CI, 1.87-2.14). Discontinued ACEI/ARB users showed the highest risk of PSE (aHR, 2.34 [95% CI, 2.15-2.54]), suggesting the importance of continuing ACEI/ARB use after stroke. Treatment-by-age interaction was significant among patients with or without ACEI/ARB use before stroke (P value for interaction 0.004 and <0.001, respectively), suggesting a stronger beneficial association in younger patients. CONCLUSIONS: The use of ACEIs/ARBs after ischemic stroke in patients with hypertension is associated with a reduced risk of PSE, especially among younger patients.

Assessment of the impact of resting heart rate on the risk of major adverse cardiovascular events after ischemic stroke: a retrospective observational study.

BACKGROUND: Although elevated heart rate is a risk factor for cardiovascular morbidity and mortality in healthy people, the association between resting heart rate and major cardiovascular risk in patients after acute ischemic stroke remains debated. This study evaluated the association between heart rate and major adverse cardiovascular events after ischemic stroke. METHODS: We conducted a retrospective cohort study analyzing data from the Chang Gung Research Database for 21,655 patients with recent ischemic stroke enrolled between January 1, 2010, and September 30, 2018. Initial in-hospital heart rates were averaged and categorized into 10-beats per minute (bpm) increments. The primary outcome was the composite of hospitalization for recurrent ischemic stroke, myocardial infarction, or all-cause mortality. Secondary outcomes were hospitalization for recurrent ischemic stroke, myocardial infarction, and heart failure. Hazard ratios and 95% confidence intervals (CIs) were estimated using Cox proportional hazards models, using the heart rate < 60 bpm subgroup as the reference. RESULTS: After a median follow-up of 3.2 years, the adjusted hazard ratios for the primary outcome were 1.13 (95% CI: 1.01 to 1.26) for heart rate 60-69 bpm, 1.35 (95% CI: 1.22 to 1.50) for heart rate 70-79 bpm, 1.64 (95% CI: 1.47 to 1.83) for heart rate 80-89 bpm, and 2.08 (95% CI: 1.85 to 2.34) for heart rate ≥ 90 bpm compared with the reference group. Heart rate ≥ 70 bpm was associated with increased risk of all secondary outcomes compared with the reference group except heart failure.  CONCLUSIONS: Heart rate is a simple measurement with important prognostic implications. In patients with ischemic stroke, initial in-hospital heart rate was associated with major adverse cardiovascular events.

Declining Incidence of Kawasaki Disease During the COVID-19 Pandemic: A Time Series Analysis.

BACKGROUND: In 2020, Taiwan experienced a dramatic disruption of daily life due to measures aimed at limiting the spread of coronavirus disease 2019 (COVID-19), although domestic COVID-19 cases were extremely rare. This provided a chance to investigate the change of incidence and possible triggers of Kawasaki disease (KD) without the interference of multisystem inflammatory syndrome in children. METHODS: This population-based study used Taiwan's National Health Insurance Research Database. To detect the change in KD incidence in children less than 6 years old during the period of social distancing policy in 2020, the autoregressive integrated moving average model was used based on seasonal variation and the increasing trend of KD incidence over the past 10 years. The trends in population mobility and common infectious diseases were also analyzed to observe their effects on KD incidence. RESULTS: The predicted monthly incidence [95% confidence interval (CI)] of KD in April, May and June 2020 were 5.26 (3.58-6.94), 5.45 (3.70-7.20) and 5.41 (3.61-7.21) per 100,000 person-months. The actual observed incidences were 2.41, 2.67 and 2.60 per 100,000 person-months, respectively, representing a 51.2%, 51.0% and 51.9% reduction in incident rates in the 3-month period of stringent COVID-19 mitigation measures. The trend was not proportionate to the profound decline of common infectious cases. CONCLUSIONS: KD incidence decreased by approximately 50% during the period of stringent COVID-19 mitigation measures in 2020 in Taiwan. Human-to-human contact may be a potential KD trigger, but there are likely other contributing factors aside from infection.

An efficient deep neural network for automatic classification of acute intracranial hemorrhages in brain CT scans.

BACKGROUND: Recent advancements in deep learning models have demonstrated their potential in the field of medical imaging, achieving remarkable performance surpassing human capabilities in tasks such as classification and segmentation. However, these modern state-of-the-art network architectures often demand substantial computational resources, which limits their practical application in resource-constrained settings. This study aims to propose an efficient diagnostic deep learning model specifically designed for the classification of intracranial hemorrhage in brain CT scans. METHOD: Our proposed model utilizes a combination of depthwise separable convolutions and a multi-receptive field mechanism to achieve a trade-off between performance and computational efficiency. The model was trained on RSNA datasets and validated on CQ500 dataset and PhysioNet dataset. RESULT: Through a comprehensive comparison with state-of-the-art models, our model achieves an average AUROC score of 0.952 on RSNA datasets and exhibits robust generalization capabilities, comparable to SE-ResNeXt, across other open datasets. Furthermore, the parameter count of our model is just 3 % of that of MobileNet V3. CONCLUSION: This study presents a diagnostic deep-learning model that is optimized for classifying intracranial hemorrhages in brain CT scans. The efficient characteristics make our proposed model highly promising for broader applications in medical settings.

Diabetic peripheral neuropathy: age-stratified glycemic control.

Background: We explore the effect of suboptimal glycemic control on the incidence of diabetic peripheral neuropathy (DPN) in both non-elderly and elderly patients with type 2 diabetes mellitus (T2DM). Methods: A 6-year follow-up study (2013-2019) enrolled T2DM patients aged >20 without DPN. Participants were classified into two groups: those below 65 years (non-elderly) and those 65 years or older (elderly). Biochemical measurements, including glycated hemoglobin (HbA1C), were recorded regularly. DPN was diagnosed using the Michigan Neuropathy Screening Instrument examination. The outcome was DPN occurrence in 2019. Results: In 552 enrollments (69% non-elderly), DPN occurred in 8.4% non-elderly and 24.0% elderly patients. A higher initial HbA1C level was significantly linked with a higher risk of future DPN in the non-elderly group (adjusted odds ratio [AOR] 1.46, 95% CI 1.13-1.89, p=0.004). In comparison, HbA1c at the end of the study period was not associated with DPN in the non-elderly group (AOR 1.17, 95% CI 0.72-1.90, p=0.526). In the elderly group, no statistical relationship was found between HbA1C levels and DPN, either in 2013 or in 2019. Conclusion: Suboptimal glycemic control at baseline, rather than at the end of the study period, predicts an increased risk of future DPN in individuals with T2DM under age 65. This correlation is not seen in elderly patients. Therefore, we recommend implementing enhanced glycemic control early in middle-aged T2DM patients and propose individualized therapeutic strategies for diabetes in different age groups.

ECG-surv: A deep learning-based model to predict time to 1-year mortality from 12-lead electrocardiogram.

BACKGROUND: Electrocardiogram (ECG) abnormalities have demonstrated potential as prognostic indicators of patient survival. However, the traditional statistical approach is constrained by structured data input, limiting its ability to fully leverage the predictive value of ECG data in prognostic modeling. METHODS: This study aims to introduce and evaluate a deep-learning model to simultaneously handle censored data and unstructured ECG data for survival analysis. We herein introduce a novel deep neural network called ECG-surv, which includes a feature extraction neural network and a time-to-event analysis neural network. The proposed model is specifically designed to predict the time to 1-year mortality by extracting and analyzing unique features from 12-lead ECG data. ECG-surv was evaluated using both an independent test set and an external set, which were collected using different ECG devices. RESULTS: The performance of ECG-surv surpassed that of the Cox proportional model, which included demographics and ECG waveform parameters, in predicting 1-year all-cause mortality, with a significantly higher concordance index (C-index) in ECG-surv than in the Cox model using both the independent test set (0.860 [95% CI: 0.859- 0.861] vs. 0.796 [95% CI: 0.791- 0.800]) and the external test set (0.813 [95% CI: 0.807- 0.814] vs. 0.764 [95% CI: 0.755- 0.770]). ECG-surv also demonstrated exceptional predictive ability for cardiovascular death (C-index of 0.891 [95% CI: 0.890- 0.893]), outperforming the Framingham risk Cox model (C-index of 0.734 [95% CI: 0.715-0.752]). CONCLUSION: ECG-surv effectively utilized unstructured ECG data in a survival analysis. It outperformed traditional statistical approaches in predicting 1-year all-cause mortality and cardiovascular death, which makes it a valuable tool for predicting patient survival.

Predicting ischemic stroke patients' prognosis changes using machine learning in a nationwide stroke registry.

Accurately predicting the prognosis of ischemic stroke patients after discharge is crucial for physicians to plan for long-term health care. Although previous studies have demonstrated that machine learning (ML) shows reasonably accurate stroke outcome predictions with limited datasets, to identify specific clinical features associated with prognosis changes after stroke that could aid physicians and patients in devising improved recovery care plans have been challenging. This study aimed to overcome these gaps by utilizing a large national stroke registry database to assess various prediction models that estimate how patients' prognosis changes over time with associated clinical factors. To properly evaluate the best predictive approaches currently available and avoid prejudice, this study employed three different prognosis prediction models including a statistical logistic regression model, commonly used clinical-based scores, and a latest high-performance ML-based XGBoost model. The study revealed that the XGBoost model outperformed other two traditional models, achieving an AUROC of 0.929 in predicting the prognosis changes of stroke patients followed for 3 months. In addition, the XGBoost model maintained remarkably high precision even when using only selected 20 most relevant clinical features compared to full clinical datasets used in the study. These selected features closely correlated with significant changes in clinical outcomes for stroke patients and showed to be effective for predicting prognosis changes after discharge, allowing physicians to make optimal decisions regarding their patients' recovery.

Polygenic Risk Score Predicts Earlier-Onset Adult Systemic Lupus Erythematosus and First-Year Renal Diseases in a Taiwanese Cohort.

OBJECTIVES: This study aimed to develop a predictive model using polygenic risk score (PRS) to forecast renal outcomes for adult systemic lupus erythematosus (SLE) in a Taiwanese population. METHODS: Patients with SLE (n=2782) and matched non-SLE controls (n=11 128) were genotyped using Genome-Wide TWB 2.0 single-nucleotide polymorphism (SNP) array. PRS models (C+T, LDpred2, Lassosum, PRSice-2, PRS-continuous shrinkage (CS)) were constructed for predicting SLE susceptibility. Logistic regression was assessed for C+T-based PRS association with renal involvement in patients with SLE. RESULTS: In the training set, C+T-based SLE-PRS, only incorporating 27 SNPs, outperformed other models with area under the curve (AUC) values of 0.629, surpassing Lassosum (AUC=0.621), PRSice-2 (AUC=0.615), LDpred2 (AUC=0.609) and PRS-CS (AUC=0.602). Additionally, C+T-based SLE-PRS demonstrated consistent predictive capacity in the testing set (AUC=0.620). Individuals in the highest quartile exhibited earlier SLE onset (39.06 vs 44.22 years, p<0.01), higher Systemic Lupus Erythematosus Disease Activity Index scores (3.00 vs 2.37, p=0.04), elevated risks of renal involvement within the first year of SLE diagnosis, including WHO class III-IV lupus nephritis (OR 2.36, 95% CI 1.47 to 3.80, p<0.01), estimated glomerular filtration rate <60 mL/min/1.73m2 (OR 1.49, 95% CI 1.18 to 1.89, p<0.01) and urine protein-to-creatinine ratio >150 mg/day (OR 2.07, 95% CI 1.49 to 2.89, p<0.01), along with increased seropositivity risks, compared with those in the lowest quartile. Furthermore, among patients with SLE with onset before 50 years, the highest PRS quartile was significantly associated with more serious renal diseases within the first year of SLE diagnosis. CONCLUSIONS: PRS of SLE is associated with earlier onset, renal involvement within the first year of SLE diagnosis and seropositivity in Taiwanese patients. Integrating PRS with clinical decision-making may enhance lupus nephritis screening and early treatment to improve renal outcomes in patients with SLE.

Antibiotic exposure during pregnancy increases risk for childhood atopic diseases: a nationwide cohort study.

PURPOSE: The prevalence of atopic diseases has increased in recent decades. A possible link between antibiotic use during pregnancy and childhood atopic disease has been proposed. The aim of this study is to explore the association of antibiotic exposure during pregnancy with childhood atopic diseases from a nationwide, population-based perspective. METHODS: This was a nationwide population-based cohort study. Taiwan's National Health Insurance Research Database was the main source of data. The pairing of mothers and children was achieved by linking the NHIRD with the Taiwan Maternal and Child Health Database. This study enrolled the first-time pregnancies from 2004 to 2010. Infants of multiple delivery, preterm delivery, and death before 5 years old were excluded. All participants were followed up at least for 5 years. Antenatal antibiotics prescribed to mothers during the pregnancy period were reviewed. Children with more than two outpatient visits, or one admission, with a main diagnosis of asthma, allergic rhinitis, or atopic dermatitis were regarded as having an atopic disease. RESULTS: A total of 900,584 children were enrolled in this study. The adjusted hazard ratios of antibiotic exposure during pregnancy to childhood atopic diseases were 1.12 for atopic dermatitis, 1.06 for asthma, and 1.08 for allergic rhinitis, all of which reached statistical significance. The trimester effect was not significant. There was a trend showing the higher the number of times a child was prenatally exposed to antibiotics, the higher the hazard ratio was for childhood atopic diseases. CONCLUSIONS: Prenatal antibiotic exposure might increase the risk of childhood atopic diseases in a dose-dependent manner.

The use of individual-based FDG-PET volume of interest in predicting conversion from mild cognitive impairment to dementia.

BACKGROUND: Based on a longitudinal cohort design, the aim of this study was to investigate whether individual-based 18F fluorodeoxyglucose positron emission tomography (18F-FDG-PET) regional signals can predict dementia conversion in patients with mild cognitive impairment (MCI). METHODS: We included 44 MCI converters (MCI-C), 38 non-converters (MCI-NC), 42 patients with Alzheimer's disease with dementia, and 40 cognitively normal controls. Data from annual cognitive measurements, 3D T1 magnetic resonance imaging (MRI) scans, and 18F-FDG-PET scans were used for outcome analysis. An individual-based FDG-PET approach was applied using seven volumes of interest (VOIs), Z transformed using a normal FDG-PET template. Hypometabolism was defined as a Z score < -2 of regional standard uptake value ratio. For the longitudinal cognitive test scores, generalized estimating equations were used. A linear mixed-effects model was used to compare the temporal impact of cortical hypometabolism and cortical thickness degeneration. RESULTS: The clinical follow-up period was 6.6 ± 3.8 years (range 3.1 to 16.0 years). The trend of cognitive decline could differentiate MCI-C from MCI-NC after 3 years of follow-up. In the baseline 18F-FDG-PET scan of the patients with MCI, medial temporal lobe (MTL; 94.7% sensitivity, 80.5% specificity) and posterior cingulate cortex (PCC; 89.5% sensitivity, 73.1% specificity) hypometabolism predicted conversion with high accuracy. 18F-FDG-PET hypometabolism preceded dementia conversion at an interval of 3.70 ± 1.68 years and was earlier than volumetric changes, with the exception of the MTL. CONCLUSIONS: Our finding supports the use of individual-based 18F-FDG-PET analysis to predict MCI conversion to dementia. Reduced FDG-PET metabolism in the MTL and PCC were strongly associated with future cognitive decline in the MCI-C group. Changes in 18F-FDG-PET occurred 1 to 8 years prior to conversion to dementia. Progressive hypometabolism in the PCC, precuneus and lateral temporal lobe, but not MTL, preceded MRI findings at the MCI stage.

Machine learning models for predicting unscheduled return visits of patients with abdominal pain at emergency department and validation during COVID-19 pandemic: A retrospective cohort study.

Machine learning (ML) models for predicting 72-hour unscheduled return visits (URVs) for patients with abdominal pain in the emergency department (ED) were developed in a previous study. This study refined the data to adjust previous prediction models and evaluated the model performance in future data validation during the COVID-19 era. We aimed to evaluate the practicality of the ML models and compare the URVs before and during the COVID-19 pandemic. We used electronic health records from Chang Gung Memorial Hospital from 2018 to 2019 as a training dataset, and various machine learning models, including logistic regression (LR), random forest (RF), extreme gradient boosting (XGB), and voting classifier (VC) were developed and subsequently used to validate against the 2020 to 2021 data. The models highlighted several determinants for 72-hour URVs, including patient age, prior ER visits, specific vital signs, and medical interventions. The LR, XGB, and VC models exhibited the same AUC of 0.71 in the testing set, whereas the VC model displayed a higher F1 score (0.21). The XGB model demonstrated the highest specificity (0.99) and precision (0.64) but the lowest sensitivity (0.01). Among these models, the VC model showed the most favorable, balanced, and comprehensive performance. Despite the promising results, the study illuminated challenges in predictive modeling, such as the unforeseen influences of global events, such as the COVID-19 pandemic. These findings not only highlight the significant potential of machine learning in augmenting emergency care but also underline the importance of iterative refinement in response to changing real-world conditions.

The impact of sodium-glucose co-transporter-2 inhibitors on dementia and cardiovascular events in diabetic patients with atrial fibrillation.

AIMS: The effectiveness of sodium-glucose co-transporter-2 inhibitors (SGLT2i) on incident dementia in patients with diabetes and atrial fibrillation (AF) remains unknown. This study aimed to investigate the association between SGLT2i and the risk of incident dementia in diabetic patients with AF, and to explore the interactions with oral anticoagulants or dipeptidyl peptidase-4 inhibitors (DPP4i). MATERIALS AND METHODS: We conducted a cohort study using Taiwan's National Health Insurance Research Database. Patients with diabetes and AFwithout a prior history of established cardiovascular diseases, were identified. Using propensity score matching, 810 patients receiving SGLT2i were matched with 1620 patients not receiving SGLT2i. The primary outcome was incident dementia, and secondary outcomes included composite cardiovascular events and mortality. RESULTS: After up to 5 years of follow-up, SGLT2i use was associated with a significantly lower risk of incident dementia (hazard: 0.71, 95% confidence interval: 0.51-0.98), particularly vascular dementia (HR: 0.44, 95% CI: 0.24-0.82). SGLT2i was related to reduced risks of AF-related hospitalisation (HR: 0.72, 95% CI: 0.56-0.93), stroke (HR: 0.75, 95% CI: 0.60-0.94), and all-cause death (HR: 0.33, 95% CI: 0.24-0.44). The protective effects were consistent irrespective of the concurrent use of non-vitamin K antagonist oral anticoagulants (NOACs) or DPP4i. CONCLUSIONS: In diabetic patients with AF, SGLT2i was associated with reduced risks of incident dementia, AF-related hospitalisation, stroke, and all-cause death. The protective effects were independent of either concurrent use of NOACs or DPP4i.

Increased NAFLD risk in newly diagnosed patients with RA during the first 4 years of follow-up: a nationwide, population-based cohort study.

BACKGROUND: Although the non-alcoholic fatty liver disease (NAFLD) is prevalent in the general population, NAFLD risk in newly diagnosed rheumatoid arthritis (RA) has rarely been explored. In this population-based cohort, we examined NAFLD risk in patients with RA and identified the potential risk factors. DESIGN: Retrospective study. SETTING: Taiwan. PARTICIPANTS: 2281 newly diagnosed patients with RA and selected 91 240 individuals without RA to match with patients with RA (1:40) by age, gender, income status and urbanisation level of the residence. OUTCOMES: In this retrospective study using the 2000-2018 claim data from two-million representative Taiwanese population, we identified and compared the incidence rates (IRs) of NAFLD and alcoholic fatty liver disease (AFLD) between RA and non-RA groups. Using multivariable regression analyses, we estimated adjusted HR (aHR) of NAFLD development in patients with RA compared with individuals without RA, with 95% CIs. RESULTS: The incidences of NALFD and AFLD were not significantly different between individuals with RA and without RA during the 17-year follow-up period. However, patients with RA had significantly increased NAFLD risk during the first 4 years after RA diagnosis, with IR ratio of 1.66 fold (95% CI 1.18 to 2.33, p<0.005), but the risk was reduced after the first 4 years. Multivariable regression analyses revealed that aHR was 2.77-fold greater in patients not receiving disease-modifying anti-rheumatic drugs therapy than in non-RA subjects (p<0.05). Old age, women, low-income status and obesity could significantly predict NAFLD development. CONCLUSIONS: We demonstrated elevated risk of NAFLD in patients with RA during the first 4 years after RA diagnosis, and old age, women, low-income status and obesity were significant predictors of NAFLD.