Chang-Fu Kuo

Aims: Sodium–glucose cotransporter‐2 inhibitors (SGLT2i), a class of oral hypoglycaemic agents (OHAs), have demonstrated cardiovascular and renal benefits. However, safety concerns may hinder their broader adoption, particularly among frail populations. We evaluated key safety outcomes associated with SGLT2i compared with dipeptidyl peptidase‐4 inhibitors (DPP4i), an OHA class with low risks of complications. Methods: We conducted a target trial emulation (TTE) study using Taiwan's National Health Insurance Research Database (NHIRD) and enrolled participants with type 2 diabetes mellitus (T2DM). Eligible participants were categorized according to newly initiated SGLT2i ( n = 224,510) or newly initiated DPP4i ( n = 698,012) with continuous use for at least 30 days. Propensity score overlap weighting (PSOW) was used to balance baseline characteristics. Results: In the intention‐to‐treat analysis, the SGLT2i group showed significantly lower incidence rate (per person‐years) of several safety outcomes than the DPP4i group, including urinary tract infection (UTI) (23.84‰ vs. 27.88‰; hazard ratio [HR], 0.86; 95% confidence interval [CI], 0.84–0.87), fractures or falls (16.55‰ vs. 17.79‰; HR, 0.93; 95% CI, 0.91–0.95), acute kidney injury (AKI) (1.06‰ vs. 1.91‰; HR, 0.55; 95% CI, 0.51–0.59), bullous pemphigoid (0.02‰ vs. 0.07‰; HR, 0.35; 95% CI, 0.23–0.54), and all‐cause mortality (13.05‰ vs. 19.72‰; HR, 0.66; 95% CI, 0.65–0.67). Conversely, the risk of genital infection was higher in the SGLT2i group (2.96‰ vs. 2.05‰; HR, 1.44; 95% CI, 1.38–1.51). These findings remained consistent in the per‐protocol analysis, across most vulnerable subgroups and individual SGLT2i and DPP4i agents. Conclusions: SGLT2i, compared with DPP4i, demonstrated a favourable safety profile across most evaluated outcomes, except for genital infections.

Objective: We examined the independent risk of incident comorbidities associated with prescribed oral paracetamol, NSAIDs and opioids and their interaction and mediation with OA. Methods: Exposure of these analgesics and the association with nine systemic groups of incident comorbidities were examined in this 5-year cohort study of 261 273 people with incident OA and 261 273 age-, sex- and practice-matched controls (no OA) in the large UK primary care database. A propensity score–adjusted time-varying exposure analysis was undertaken using multivariable Cox models to estimate the hazard ratios (HRs) for incident comorbidities associated with OA and analgesics. Interaction was defined when the combined term for OA and analgesic was significant. The proportion of the risk associated with OA mediated by analgesics was also calculated. Results: The mean age of the cohort was 60 years and 57.7% were female. NSAIDs and opioids were independently associated with all comorbidities. Paracetamol interacted with OA leading to an extra risk of renal comorbidities [multiplicative HR (mHR) 1.14 (95% CI 1.06, 1.22)]. NSAIDs and opioids interacted with OA with extra risk of cardiovascular [mHR 1.05 (95% CI 1.01, 1.11) and mHR 1.06 (95% CI 1.02, 1.11)] and endocrine [mHR 1.11 (95% CI 1.05, 1.17) and mHR 1.08 (95% CI 1.02, 1.13)] comorbidities. About 16% and 7% risk from OA to psychological comorbidities were indirectly mediated by NSAIDs and opioids, respectively. Conclusion: Our findings suggest that in people with OA, analgesic prescriptions may be associated with increased risk of certain incident comorbidities, particularly in long-term use.

Background: Tibial plateau fractures represent a diverse group of intra-articular injuries that can be difficult to detect and characterize on initial imaging. The aim of the present study was to develop an artificial intelligence (AI) diagnostic tool for identifying tibial plateau fractures on radiographs. Methods: In this retrospective study, we analyzed radiographs that had been made from January 2018 to December 2020 for 1,809 patients, with an equal distribution of male and female adults. A total of 3,821 anteroposterior and lateral knee radiographs were evaluated with use of the EfficientNet B3 AI model, with computed tomography (CT) images being used as the ground truth. Evaluation metrics focused on the area under the receiver operating characteristic curve (AUC) and positive predictive values across different subgroups. Results: Our AI model attained AUCs of 0.98 and 0.97 for detecting tibial plateau fractures in the test and external validation datasets, respectively. Subgroup analysis revealed diverse positive predictive values across different Schatzker types and 3-column classifications. Conclusions: Our deep learning model exhibits newfound ability for identifying tibial plateau fractures. However, we encountered several limitations, such as imbalances among the sizes of various subgroups in the dataset and an inability to identify radiographs containing foreign objects or other defects. Level of Evidence: Diagnostic Level III . See Instructions for Authors for a complete description of levels of evidence.

Background: Several artificial intelligence-enhanced electrocardiogram (AI-ECG) models have shown promise in detecting left ventricular systolic dysfunction (LVSD), but their head-to-head agreement and performance have not been independently compared within the same cohort. Objectives: To compare the performance of published AI-ECG models for LVSD detection in a standardized external cohort and evaluate the field’s transparency and reproducibility. Methods: We systematically reviewed AI-ECG models predicting LVSD and assessed the risk of bias. Authors were invited to share models for external validation in a well-phenotyped registry of patients undergoing routine clinical cardiac magnetic resonance imaging (CMR) with cardiologist-adjudicated reports and paired ECGs. Model performance was evaluated in all consecutive patients and a lower-complexity subgroup with 15% LVSD prevalence. Results: We identified 35 studies describing 51 models, reporting high (AUROC >0.80) or excellent (AUROC >0.90) performance. The risk of bias is high and primarily attributed to the limited description of development and validation cohort characteristics, as well as the lack of independent external validation. Four groups (from Korea, the United States, Taiwan, and the Netherlands) shared models for independent testing. AUROCs ranged from 0.83 to 0.93 in all patients (n = 1,203; mean age 59 ± 15 years; 450 [35%] female) and from 0.87 to 0.96 in the lower complexity subset. Performance remained consistent across subgroups, with slight decreases in ECGs showing wide QRS complexes or atrial fibrillation. Conclusions: In this first-in-kind independent validation and head-to-head comparison study, AI-ECG for LVSD detection demonstrated strong performance despite training on disparate populations. However, the limited availability of models hinders independent validation.

Incorporating pharmacogenetics into clinical practice promises to improve therapeutic outcomes by optimizing drug selection and dosage based on genetic factors affecting drug response. A key advantage of PGx-guided therapy is to decrease the likelihood of adverse events. To evaluate the clinical impact of PGx risk variants, we performed a retrospective study using genetic and clinical data from the largest Han Chinese cohort, comprising 486,956 individuals, assembled by the Taiwan Precision Medicine Initiative. We found that nearly all participants carried at least one genetic variant that could affect drug response, with many carrying multiple risk variants. Here we show the detailed analyses of four gene-drug pairs, azathioprine ( NUDT15 / TPMT ), clopidogrel ( CYP2C19 ), statins ( ABCG2 / CYP2C9 / SLCO1B1 ), and NSAIDs ( CYP2C9 ), for which sufficient data exists for statistical power. While the results validate previous findings that PGx risk variants are significantly associated with drug-related adverse events or ineffectiveness, the excess risk of adverse events or lack of efficacy is small compared to that found in those without the PGx risk variants, and most patients with PGx variants do not suffer from adverse events. Our results point to the complexity of implementing PGx in clinical practice and the need for integrative approaches to optimize precision medicine.

Rheumatoid arthritis (RA) is distinguished by the presence of modified bone microarchitecture, also known as ‘texture,’ in the periarticular regions. The radiographic detection of such alterations in RA can be challenging. To train and to validate a deep learning model to quantitatively produce periarticular texture features directly from radiography and predict the diagnosis of early RA without human reading. Two kinds of deep learning models were compared for diagnostic performance. Anterior–posterior bilateral hands radiographs of 891 early RA (within one year of initial diagnosis) and 1237 non-RA patients were split into a training set (64%), a validation set (16%), and a test set (20%). The second, third, and fourth distal metacarpal areas were segmented for the Deep Texture Encoding Network (Deep-TEN; texture-based) and residual network-50 (ResNet-50; texture and structure-based) models to predict the probability of RA. The area under the curve of the receiver operating characteristics curve for RA was 0.69 for the Deep-TEN model and 0.73 for the ResNet-50 model. The positive predictive values of a high texture score to classify RA using the Deep-TEN and ResNet-50 models were 0.64 and 0.67, respectively. High mean texture scores were associated with age- and sex-adjusted odds ratios (ORs) with 95% confidence interval (CI) for RA of 3.42 (2.59–4.50) and 4.30 (3.26–5.69) using the Deep-TEN and ResNet-50 models, respectively. The moderate and high RA risk groups determined by the Deep-TEN model were associated with adjusted ORs (95% CIs) of 2.48 (1.78–3.47) and 4.39 (3.11–6.20) for RA, respectively, and those using the ResNet-50 model were 2.17 (1.55–3.04) and 6.91 (4.83–9.90), respectively. Fully automated quantitative assessment for periarticular texture by deep learning models can help in the classification of early RA.