EClinicalMedicine最新文献

Background: Large electronic databases are powerful tools for studying rare diseases, however accurate Interstitial Lung Disease (ILD) classification remains challenging. Rule-based approaches rely heavily on diagnostic codes-unreliable markers of ILD. We aimed to develop and externally validate an ILD classification algorithm that robustly identifies prevalent cases using routinely captured electronic health record (EHR) data.

Methods: In this retrospective model development and validation study, we used EHR data from the UC Health Data Warehouse, a multi-institutional dataset from six academic centres in California, USA (2012-2024). Data from individuals ≥18 years with ≥ five encounters were included. We developed the Universal ILD Classifier, a machine learning model developed on standardised EHR data elements from UC San Francisco (January 1, 1981-January 6, 2025). The algorithm was converted to an EHR-agnostic common data model, to enable external validation across three independent sites (UC Irvine, Los Angeles, and San Diego; January 1, 2012-April 30, 2025). Features included diagnostic and procedure codes, laboratory results, medications, demographics, and utilisation metrics. The main outcome was algorithm performance assessed by positive predictive value (PPV), sensitivity, F1-score, and receiver operative characteristic-area under the curve (ROC-AUC). Performance was compared with two widely used rule-based classification methods.

Findings: The Universal ILD Classifier, developed on data from 203,976 patients and validated on data at three independent sites (N = 250 per site), demonstrated robust generalisability, achieving average PPV = 0.67 (0.58-0.72), sensitivity = 0.97 (0.94-0.99), F1-score = 0.79 (0.72-0.84), and ROC-AUC = 0.96 (0.94-0.97). It consistently outperformed both rule-based methods, which had PPVs = 0.55 (0.50-0.59) and 0.67 (0.61-0.73), sensitivities = 0.98 (0.96-0.99) and 0.59 (0.53-0.64), F1-scores = 0.71 (0.66-0.74) and 0.63 (0.57-0.68), and ROC-AUCs = 0.80 (0.78-0.82) and 0.73 (0.70-0.76) respectively.

Interpretation: Accurate patient identification is essential for epidemiological studies and ILD clinical trials. The Universal ILD Classifier leverages commonly available EHR data and outperforms rule-based approaches, supporting more reliable large-scale ILD research and offering a foundation for further refinement with additional features. Limitations inherent to retrospective EHR analyses, including misclassification, residual confounding, and limited generalisability, may have influenced effect estimates and should be considered when interpreting these findings.

Funding: Boehringer Ingelheim Pharmaceuticals, Inc. (BIPI).

Background: Dapiglutide, a dual glucagon-like peptide (GLP)-1 and GLP-2 receptor agonist, is under clinical development for bodyweight reduction in obesity, capitalising on GLP-1 receptor-mediated food intake-reducing effect leading to bodyweight loss and GLP-2 receptor-mediated improvement of gut barrier function and anti-inflammatory properties.

Methods: In this 12-week, investigator-initiated, phase IIa, proof-of-concept, double-blind, placebo-controlled trial in Denmark, adults (18-75 years) with obesity (BMI ≥30 kg/m²) were randomised to once-weekly subcutaneous dapiglutide (4 or 6 mg) or placebo (1:1:1; stratified by sex) without concurrent lifestyle intervention. Exclusion criteria included HbA_1c ≥48 mmol/mol and recent ≥5% bodyweight change. The primary endpoint was percentage bodyweight change from baseline to week 12. The effects of 6 mg and 4 mg dapiglutide were tested hierarchically against first placebo, next each other in efficacy analyses with missing data replaced by standard multiple imputation. The study is registered with EU (trial no. 2022-501649-54-00) and ClinicalTrials.gov (NCT05788601).

Findings: Between April 27, 2023, and April 24, 2024, 54 adults living with obesity (63% women; mean bodyweight 101·3 kg; mean BMI 35·2 kg/m²) were randomised. In the primary efficacy analysis, 6 mg dapiglutide led to a mean bodyweight change of -2·1% [95% CI -4·3 to 0·2; p = 0·076] compared to placebo. Dapiglutide appeared safe and well-tolerated, with common adverse events including reduced appetite and nausea. No participants discontinued due to drug-induced adverse events, and in general, dropout rates were low, i.e. 0% (placebo), 11% (4 mg), and 6% (6 mg).

Interpretation: Dapiglutide (4 mg or 6 mg) for 12 weeks did not result in a statistically significant difference in bodyweight change compared with placebo in persons with obesity. The safety profile was favourable, and the findings support further investigation of higher doses in larger and longer studies to evaluate the weight-lowering potential of dapiglutide.

Funding: Unrestricted grant from Zealand Pharma.

Background: Solid organ transplantation is the definitive treatment for end-stage organ failure, yet access is highly inequitable worldwide. Comparable long-term evidence across organs, regions, and development settings remains limited.

Methods: Data from the WHO Global Observatory on Donation and Transplantation (GODT) for six solid organs (2008-2023) were analyzed. Per-million population (PMP) rates and the estimated annual percentage change (EAPC) were calculated; disparities by Human Development Index (HDI), WHO regions, and Global Burden of Disease 2021 (GBD 2021) regions were quantified using the slope index of inequality (SII) and concentration index (CI); and transplant capacity gap was estimated by comparing observed volumes with PMP benchmarks from very-high-HDI countries.

Findings: Global transplants rose 76% (101,990-179,091); PMP increased 15.1 → 23.1 (EAPC 2.5%). Kidney transplantation accounted for 65% of all solid organ transplants in 2023, representing the largest share of global activity; lung transplantation showed the fastest relative growth. Two procedures declined globally-pancreas-only transplantation and small-bowel transplantation. Absolute volumes were highest in the USA, China, and India, but PMP ranged from >120 (Spain, USA) to <5 in most low-HDI countries. Growth accrued mainly in very-high-HDI settings, with minimal contribution from low-HDI regions. Japan showed persistently low rates despite very-high-HDI status, whereas Mongolia achieved the world's highest EAPC despite low HDI. Inequality widened by SII (55.9 → 73.9), while CI fell modestly (0.61 → 0.53). Benchmarking indicated the largest transplant capacity gap for kidney (>200,000 procedures), then liver (>80,000) and heart (>30,000); coverage remained <10% in most low-HDI countries.

Interpretation: Global activity increased substantially but gains concentrated in very-high-HDI countries, and inequities persist. Outlier trajectories highlight sociocultural and policy factors beyond economic development. Large benchmark-based gaps-especially for kidney, liver, and heart-remain across low- and middle-HDI settings. Strategic investment in policy, infrastructure, and integration of transplantation within universal health coverage is essential to advance equitable access.

Funding: The Integrated Fund4150102990-58803-0/Prof. Dr. Z.K. Chen; The Fund of Guangzhou Key Laboratory of Organ Transplantation2025A03J4036.

Background: Many existing randomised controlled trials lack sufficient power to assess primary kidney outcomes. This study aimed to evaluate whether statin therapy offers a clinically meaningful reno-protective effect in patients with chronic kidney disease (CKD).

Methods: In this retrospective cohort study, electronic health records in Hong Kong were extracted to perform sequential target trial emulation. Eligible adults (aged 18+ years) with CKD who met the indication for statin initiation between Jan 1, 2008 and Dec 31, 2017 were included; those with history of estimated glomerular filtration rate (eGFR) < 15 mL/min/1.73 m² were excluded. Participants were categorised as statin initiators or non-initiators at each calendar month during inclusion period, where statin initiators were propensity score-matched with non-initiators. Follow-up data were collected for all participants until the occurrence of outcomes, death, loss to follow-up (2 years after last records), or the end of data availability (Dec 31, 2022), whichever occurred first. The hazard ratio (HR) of all-cause mortality, eGFR deterioration (eGFR <15 mL/min/1.73 m², ≥30% eGFR decline, and ≥50% eGFR decline) and composite outcomes (all-cause mortality, eGFR <15 mL/min/1.73 m², and ≥50% eGFR decline) was estimated by pooled logistic regression using intention-to-treat (ITT) and per-protocol (PP) approach.

Findings: 1,437,014 eligible person-trials were identified (statin initiators n = 30,907; non-initiators n = 1,406,107), from which 30,892 statin initiators and 108,380 non-initiators were included after propensity-score matching. Relative to non-initiators, significant risk reduction was found among statin initiators in all-cause mortality (HR [95% confidence interval (CI)], ITT: 0.97 [0.95-0.98]; PP: 0.91 [0.88-0.93]), progression to eGFR <15 mL/min/1.73 m² (ITT: 0.91 [0.89-0.93]; PP: 0.77 [0.74-0.80]), ≥50% eGFR decline (ITT: 0.95 [0.93-0.98]; PP: 0.89 [0.84-0.93]), and composite outcomes (ITT: 0.96 [0.94-0.97]; PP: 0.90 [0.88-0.92]). Statin therapy initiation was also associated significantly with reduced risk of ≥30% eGFR decline using PP approach (0.94 [0.92-0.96]).

Interpretation: Over a 10-year follow-up period, initiating statin therapy in patients with CKD was associated with a small yet significant decrease in all-cause mortality and a modest reno-protective effect. Future research should aim to clarify the effects of statin intensity, duration, and adherence.

Funding: National Natural Science Foundation of China.

[This corrects the article DOI: 10.1016/j.eclinm.2025.103714.].

Background: Molecular point-of-care testing (mPOCT) offers rapid identification of respiratory pathogens, but its impact on antibiotic use and patient outcomes remains uncertain. We aimed to comprehensively evaluate the effects of mPOCT on antibiotic use and major clinical outcomes in patients presenting with acute respiratory tract infections (ARTIs).

Methods: We searched MEDLINE, Embase, Web of Science, CENTRAL, CNKI, and Wanfang Data from inception to July 1, 2025, for randomised controlled trials (RCTs) evaluating mPOCT for patients presenting with ARTIs (PROSPERO CRD420251069333). The primary outcome was antibiotic use, assessed using pooled risk ratio (RR) with random-effects models. Risk of bias and certainty of evidence were assessed using the Risk Of Bias instrument for Use in SysTematic reviews-for Randomised Controlled Trials (ROBUST-RCT) and core Grading of Recommendations, Assessment, Development and Evaluation (GRADE), respectively.

Findings: We included 25 RCTs involving 12,638 patients, of whom 61.0% were adults. Overall, mPOCT probably had little to no important effect on antibiotic use (RR 0.95, 95% CI 0.90-1.00; moderate certainty) or treatment duration (mean difference -0.44 days, 95% CI -0.98 to 0.09; moderate certainty). In adults, high-certainty evidence showed no effect on antibiotic use (RR 1.00, 95% CI 0.98-1.02), whereas in children, low-certainty evidence suggested a potential reduction (RR 0.79, 95% CI 0.65-0.97). Although mPOCT increased appropriate antibiotic prescribing (RR 2.07, 95% CI 1.55-2.77; moderate certainty), it did not affect 30-day mortality (RR 0.97, 95% CI 0.82-1.15; high certainty) and intensive care unit admission (RR 0.90, 95% CI 0.65-1.25; high certainty).

Interpretation: Moderate to high certainty evidence suggests that mPOCT does not meaningfully reduce overall antibiotic use or improve patient outcomes, particularly in adults, despite enhancing prescribing appropriateness. Routine use of mPOCT for adults with ARTIs is therefore not supported.

Funding: National Natural Science Foundation of China, the Postdoctoral Science Foundation, the Chongqing Municipality Joint Science and Health Major Medical Research Project, Outstanding Youth in Science and Technology, the Chongqing Youth Talent Fund, and the Research Foundation Flanders.

Background: Meningiomas are the most common primary brain tumours in adults. Concerns have emerged about a possible link between progestogen use and intracranial meningioma; we assessed this association.

Methods: In this systematic review and meta-analysis, we searched PubMed/MEDLINE, Embase, Cochrane Library, EPI-PHARE database (from inception up to November 01, 2025), pharmacovigilance reports, and backward snowballing. Eligible publications were English or French epidemiological studies, reporting associations between progestogens and intracranial meningiomas. We excluded non-original reports, abstracts-only, and studies without eligible progestogen exposure or meningioma outcomes. We extracted summary data from published reports. Risk of bias was assessed with the Newcastle-Ottawa Scale, and certainty of evidence with GRADE. The primary outcome was intracranial meningioma. Secondary outcomes were malignancy, location, and regression. Random-effects models were used, and heterogeneity was assessed with I²; a narrative synthesis was also performed.

Findings: Of 542 records screened, 78 studies were included in the review, and 14 high-quality observational studies in meta-analysis; all 14 were NOS high quality, although residual confounding and potential outcome misclassification cannot be excluded. Cyproterone acetate (CPA) was associated with increased meningioma risk (5 studies; 1047 exposed; pooled-OR 12.36 (95% CI: 7.47-20.45); I²: 73.8%; GRADE: moderate). Depot medroxyprogesterone acetate was also associated (6 studies; 842 exposed; pooled-OR 2.68 (95% CI: 1.72-4.19); I²: 92.7%; GRADE: low). Chlormadinone acetate (CMA), nomegestrol acetate (NOMAC), promegestone, medrogestone, and desogestrel showed signals of increased risk (CMA 3 studies, 164-683 exposed; NOMAC 3, 171-969; promegestone 1, 83; medrogestone 1, 42; desogestrel 2, 115-287). We did not pool these estimates due to sparse, heterogeneous evidence. No signal was found for norgestrel, levonorgestrel, progesterone, dydrogesterone, or spironolactone; evidence for dienogest and hydroxyprogesterone was insufficient. Regression after withdrawal was reported for CPA and NOMAC. Tumours were predominantly anterior/middle skull base, and malignant meningiomas were more frequent with CPA, CMA, and NOMAC.

Interpretation: The certainty of evidence was limited by the observational design, residual confounding, heterogeneity, and imprecision for some exposures. Use of specific progestogens, particularly high dose macroprogestogens may be associated with an increased risk of intracranial meningioma. Transparent patient information and careful clinical and, where appropriate, imaging follow-up are essential.

Funding: None.

Digital health technologies and artificial intelligence (AI), are transforming medical research, health care, and public health. The ever-increasing usage of algorithms in health care has challenged governments, regulatory agencies, health organizations, developers, and providers, and AI raises novel ethical challenges that extend beyond the jurisdiction of traditional borders and regulatory health-care processes and structures. While there is growing consensus in recognizing these ethical challenges, there is less agreement over the necessary AI guardrails. This Viewpoint offers a synthesis of representative AI-enabled health policy approaches across jurisdictions and advances practical recommendations for an adaptive, international AI policy and governance framework that will be responsible for monitoring and advancing its regulations in pace with the rapid growth of AI technologies. We use the Declaration of Helsinki as a normative reference point, to derive risk-proportionate safeguards for AI-enabled health across research and non-research settings.

Background: Objective assessment of left ventricular function remains a key prognosticator that is used to guide therapeutic decisions for patients with heart failure (HF). However, the left ventricular ejection fraction (LVEF) is dynamic, with worsening LVEF linked to increased morbidity and mortality. Identifying patients at risk of LVEF decline would improve prognostication and enable timely therapeutic intervention.

Methods: We developed a deep learning model to Predict changes in left ventricULar Systolic function from Electrocardiograms (ECG) of patients who have Heart Failure (PULSE-HF). The model integrates 12-lead ECG waveforms with a patient's history of prior LVEF measurements to calculate the likelihood that the LVEF will be less than 40% during the year after the ECG is obtained. The model is retrospectively developed and tested using data from one hospital and externally validated on retrospective cohorts from two different hospitals. The internal development data was collected between January 1, 2000, and June 30, 2021. The external validation datasets were collected between January 1, 2000, and June 30, 2021 at one hospital and between 2008 and 2019 at the other hospital.

Findings: PULSE-HF demonstrates strong discriminatory ability with respect to forecasting whether the LVEF would be below 40% within the next year, achieving areas under the receiver operating characteristic curve (AUROC) of 87.5-91.4% across all three HF cohorts. Among patients with HF who have a baseline LVEF above 40%, PULSE-HF effectively identified those at risk of worsening LVEF with AUROCs of 81.6-86.3% across all three datasets. PULSE-HF's discriminatory ability remained consistently high across a range of subgroups with different comorbidities and regardless of medical therapy. Assuming an underlying prevalence of LVEF worsening of 10% per year, PULSE-HF's negative predictive values are over 97%, assuming an underlying sensitivity of 80%. Lastly, we demonstrate that a lead I version of PULSE-HF has a performance similar to the performance of the model that uses all 12 ECG leads.

Interpretation: PULSE-HF robustly predicts worsening LVEF in patients who have a prior diagnosis of HF. The method provides a platform for identifying patients who are at an increased risk of worsening systolic dysfunction.

Funding: This work was supported, in part, by a grant from Quanta Computers.