Journal of Thrombosis and Haemostasis最新文献

{"title":"Pharmacokinetic-guided dose reduction of emicizumab in congenital hemophilia A: interim analysis of the DosEmi study","authors":"Konrad van der Zwet ,&nbsp;Amber D.W. de Vos ,&nbsp;Marjon H. Cnossen ,&nbsp;Floor C.J.I. Heubel-Moenen ,&nbsp;Saskia E.M. Schols ,&nbsp;Paula F. Ypma ,&nbsp;Paul L. den Exter ,&nbsp;Hélène L. Hooimeijer ,&nbsp;Michiel Coppens ,&nbsp;Ron A.A. Mathôt ,&nbsp;Alexander Janssen ,&nbsp;Anouk A.M.T. Donners ,&nbsp;Ilmar Kruis ,&nbsp;Rolf T. Urbanus ,&nbsp;Lize F.D. van Vulpen ,&nbsp;Corien L. Eckhardt ,&nbsp;Roger E.G. Schutgens ,&nbsp;Kathelijn Fischer","doi":"10.1016/j.jtha.2025.10.029","DOIUrl":"10.1016/j.jtha.2025.10.029","url":null,"abstract":"<div><h3>Background</h3><div>Emicizumab provides effective prophylaxis for hemophilia A (HA), but its cost puts a burden on health care systems. Pharmacokinetic (PK)-pharmacodynamic analyses suggest that emicizumab is effective at lower trough concentrations (Ctrough) than the mean 55 μg/mL reported with conventional dosing.</div></div><div><h3>Objectives</h3><div>To perform an interim analysis comparing bleeding control during 6 months of conventional emicizumab dosing vs PK-guided dosing in participants with HA, targeting a Ctrough of 25-35 μg/mL.</div></div><div><h3>Methods</h3><div>The DosEmi study (NCT06320626) is an ongoing multicenter, open-label, crossover Dutch trial. Eligible participants were aged ≥16 years using emicizumab for ≥12 months with good bleeding control (≤2 treated bleeds/6 months, none spontaneous). This planned interim analysis compared 6 months’ bleeding on conventional dosing with PK-guided dosing. Study continuation criteria were ≤15% decrease in the proportion of participants without treated bleeds, maximum of 2 spontaneous bleeds overall, and &lt;1.0 increase in bleeds/y (annualized bleeding rate).</div></div><div><h3>Results</h3><div>In 26 participants with severe HA, PK-guided dosing reduced emicizumab consumption by 39% (range, 13%-50%). The proportion of participants without treated bleeds was 69% with conventional dosing vs 58% with PK-guided dosing (risk difference, +11%; <em>P</em> = .254); the proportion without joint bleeds remained stable at 85% vs 88%, respectively (risk difference, −4%; <em>P</em> = .500). The annualized bleeding rate remained low (0.7 vs 0.9; <em>P</em> = .132), including 1 spontaneous muscle bleed (emicizumab concentration 52.5 μg/mL).</div></div><div><h3>Conclusion</h3><div>These interim results suggest that bleeding control with PK-guided reduced emicizumab dosing (target Ctrough 25-35 μg/mL) is similar to conventional dosing. All prespecified criteria for study continuation and inclusion of participants aged &lt;16 years were met.</div></div>","PeriodicalId":17326,"journal":{"name":"Journal of Thrombosis and Haemostasis","volume":"24 3","pages":"Pages 890-899"},"PeriodicalIF":5.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145481886","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Association of direct oral anticoagulant–anticonvulsant coprescription with clinical outcomes in older adults: a population-based cohort study","authors":"Stephanie Carlin ,&nbsp;Anne M. Holbrook ,&nbsp;Matteo Candeloro ,&nbsp;Francis Nguyen ,&nbsp;J. Michael Paterson ,&nbsp;James Douketis","doi":"10.1016/j.jtha.2025.10.027","DOIUrl":"10.1016/j.jtha.2025.10.027","url":null,"abstract":"<div><h3>Background</h3><div>Anticonvulsants vary in their theoretical propensity to interact with direct oral anticoagulants (DOACs). Coprescription may reduce DOAC efficacy and increase risk of thromboembolism.</div></div><div><h3>Objectives</h3><div>To evaluate the risk of clinical outcomes - thromboembolism, major bleeding, and death - among patients who were receiving a DOAC and newly treated with an interacting anticonvulsant relative to those newly treated with a non-interacting anticonvulsant.</div></div><div><h3>Methods</h3><div>We undertook 3 cohort studies using administrative health care data for Ontarians aged ≥66 years who were using a DOAC and newly prescribed 1 of 3 groups of anticonvulsants with declining theoretic propensity for interaction: group 1 (strongly interacting); group 2 (mildly interacting); and group 3 (potentially interacting). Each cohort was compared with a reference group of patients who were newly coprescribed a DOAC with a presumed noninteracting anticonvulsant. The primary outcome was hospitalization for thromboembolism. Secondary outcomes were major bleeding, death, and a composite of thromboembolism or death. Analysis was by propensity score inverse probability of treatment weighted competing risk Fine-Gray regression models.</div></div><div><h3>Results</h3><div>We studied 17 325 patients: 878 in group 1; 313 in group 2; 943 in group 3; and 15 191 in the reference group. After inverse probability of treatment weighted, we did not observe an association with thromboembolism or major bleeding in groups 1 to 3. The Cox proportional hazards ratio for death in those prescribed group 1 anticonvulsants was 2.21 (95% CI, 1.77-2.75).</div></div><div><h3>Conclusion</h3><div>In patients using a DOAC and newly prescribed a group 1, 2, or 3 anticonvulsant, we did not observe an increased risk of thromboembolism. In patients newly coprescribed a DOAC with a group 1 anticonvulsant, the observed increased risk of death is likely due to residual confounding.</div></div>","PeriodicalId":17326,"journal":{"name":"Journal of Thrombosis and Haemostasis","volume":"24 3","pages":"Pages 1000-1013"},"PeriodicalIF":5.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145481818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Historical review: more than two decades understanding the genetic architecture of hemostasis and thrombosis","authors":"Maria Sabater-Lleal ,&nbsp;Florian Thibord ,&nbsp;Paul S. de Vries ,&nbsp;Jennifer Huffman ,&nbsp;Alisa S. Wolberg ,&nbsp;Charles J. Lowenstein ,&nbsp;Alanna C. Morrison ,&nbsp;Andrew D. Johnson ,&nbsp;Nicholas L. Smith","doi":"10.1016/j.jtha.2025.11.006","DOIUrl":"10.1016/j.jtha.2025.11.006","url":null,"abstract":"<div><div>From the beginning of the millennium and the development of genome-wide analyses, the technical advances and remarkable increase in research sample sizes have led to an escalating number of discoveries revealing genetic determinants of levels of the main factors regulating hemostasis and thrombosis and demonstrating a clear polygenic complex regulation of most coagulation factors. These discoveries have been useful to understand the biology underlying hemostasis regulation and to understand risk of associated thrombotic disease, such as venous thromboembolism, coronary artery disease, and ischemic stroke. In this historical review, we outline the main discoveries in genetic studies of coagulation factors (fibrinogen and its alternatively spliced γ' isoform, D-dimer, factor [F]V, FVII, FVIII, von Willebrand factor, and FXI), the main natural anticoagulants (protein C, protein S, and antithrombin), components of fibrinolysis (tissue plasminogen activator and plasminogen activator inhibitor-1), and global coagulation tests (prothrombin time and activated partial thromboplastin time). We explore the clinical implications of these discoveries and suggest new avenues for future investigation.</div></div>","PeriodicalId":17326,"journal":{"name":"Journal of Thrombosis and Haemostasis","volume":"24 3","pages":"Pages 1104-1117"},"PeriodicalIF":5.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145696274","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting treatment-free remission after thrombopoietin receptor agonist therapy in immune thrombocytopenia: real-world outcomes from a Korean cohort","authors":"Jae-Ho Yoon,&nbsp;Seonghan Lee,&nbsp;Daehun Kwag,&nbsp;Gi June Min,&nbsp;Sung-Soo Park,&nbsp;Silvia Park,&nbsp;Sung-Eun Lee,&nbsp;Byung-Sik Cho,&nbsp;Ki-Seong Eom,&nbsp;Yoo-Jin Kim,&nbsp;Hee-Je Kim,&nbsp;Chang-Ki Min,&nbsp;Seok-Goo Cho","doi":"10.1016/j.jtha.2025.11.017","DOIUrl":"10.1016/j.jtha.2025.11.017","url":null,"abstract":"<div><h3>Background</h3><div>Thrombopoietin receptor agonists (TPO-RAs) have become a cornerstone in the management of relapsed or refractory immune thrombocytopenia (R/R-ITP), with emerging evidence supporting treatment discontinuation in selected patients. However, predictors of sustained response and real-world switching patterns remain underexplored in Asian populations.</div></div><div><h3>Objectives</h3><div>To evaluate clinical outcomes, switching strategies, and predictors of treatment-free remission following TPO-RA therapy in Korean patients with R/R-ITP.</div></div><div><h3>Methods</h3><div>We retrospectively analyzed 91 adult R/R-ITP patients treated with eltrombopag, romiplostim (ROMI), or both at a tertiary center in Korea from 2016 to 2021. Clinical responses, platelet count trajectories, loss of response, and sustained response off treatment (SROT) after TPO-RA were assessed. Subgroup analyses identified predictors for response and SROT rates. Notably, Korean reimbursement guidelines mandating drug interruption every 6 months facilitated longitudinal assessment of treatment-free outcomes.</div></div><div><h3>Results</h3><div>The overall response rate was 93.4%, and 40% of those achieved SROT after TPO-RA. Higher peak platelet count (platelet count at maximal response &gt; 100 × 10⁹/L; odds ratio, 5.28; <em>P</em> = .005) and prior intravenous immunoglobulin (IVIG) responsiveness (odds ratio, 3.93; <em>P</em> = .036) were significantly associated with SROT. Despite a significantly higher platelet count at maximal response with ROMI, response and SROT rates were comparable between eltrombopag and ROMI, likely reflecting differences in dosing and adherence patterns. Patients with high autoimmune burden or poor IVIG response showed higher loss of response and lower SROT rates.</div></div><div><h3>Conclusion</h3><div>TPO-RAs demonstrated robust efficacy and potential for SROT in R/R-ITP. Baseline immunologic status and prior IVIG response may guide treatment selection and discontinuation strategies. These findings highlight the importance of individualized immune thrombocytopenia management and suggest future directions for patients with persistent refractoriness.</div></div>","PeriodicalId":17326,"journal":{"name":"Journal of Thrombosis and Haemostasis","volume":"24 3","pages":"Pages 1130-1140"},"PeriodicalIF":5.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145775019","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mechanical heart valves: an idea from the past looking to the future","authors":"Giuseppe Santarpino ,&nbsp;Chiara Maria Prencipe ,&nbsp;Veronica D’Anna","doi":"10.1016/j.jtha.2025.11.032","DOIUrl":"10.1016/j.jtha.2025.11.032","url":null,"abstract":"","PeriodicalId":17326,"journal":{"name":"Journal of Thrombosis and Haemostasis","volume":"24 3","pages":"Pages 858-860"},"PeriodicalIF":5.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147321960","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Intraosseous administration of lyophilized synthetic platelets renders hemostatic efficacy in rat model of traumatic hemorrhage","authors":"Zeyu Liu ,&nbsp;Abiha Abdullah ,&nbsp;Zachary A. Secunda ,&nbsp;Norman F. Luc ,&nbsp;Lauren Jackson ,&nbsp;Baylee Traylor ,&nbsp;Dante Disharoon ,&nbsp;Christa Pawlowski ,&nbsp;Riya Sharma ,&nbsp;Ujjal Didar Singh Sekhon ,&nbsp;Emma Quill ,&nbsp;Philip Charles Spinella ,&nbsp;Susan M. Shea ,&nbsp;Michael A. Bruckman ,&nbsp;Anirban Sen Gupta ,&nbsp;Matthew D. Neal","doi":"10.1016/j.jtha.2025.10.030","DOIUrl":"10.1016/j.jtha.2025.10.030","url":null,"abstract":"<div><h3>Background</h3><div>In traumatic hemorrhage, transfusion of donor-derived platelets improve hemostasis and survival, but their availability is often limited by supply constraints and short shelf-life. To address this, we have developed SynthoPlate (SP), a synthetic platelet nanotechnology that mimics the primary hemostatic functions of natural platelets, and have recently advanced its manufacturing into a shelf-stable, lyophilized powder for rapid aqueous-reconstitution and on-demand use.</div></div><div><h3>Objectives</h3><div>We aimed to evaluate the feasibility, safety, and efficacy of administering SP intraosseously in a rat model of traumatic hemorrhage, considering the critical clinical relevance of intraosseous access in prehospital and combat medicine.</div></div><div><h3>Methods</h3><div>We first assessed SP’s hemostatic cooperativity with rat platelets using a microfluidic assay. Next, SP was administered intraosseously in rats to evaluate safety and biodistribution. Finally, 0.5 mg/kg of SP was administered intraosseously in a rat liver laceration model to assess effects on hemodynamics, blood loss, and survival.</div></div><div><h3>Results</h3><div>Microfluidic studies confirmed SP’s hemostatic capability in platelet-depleted rat plasma. In pilot safety studies, intraosseously administered SP was well tolerated at doses up to 20 mg/kg, 40 times the proposed effective dose. In efficacy studies, rats treated with SP showed significantly reduced blood loss and improved survival compared with controls. SP-treated rats also exhibited higher mean arterial pressure postinjury, shorter durations of hypotension, and faster mean arterial pressure recovery.</div></div><div><h3>Conclusion</h3><div>This first-in-kind study demonstrates the feasibility of administering SP intraosseously to enhance hemostasis and survival in traumatic hemorrhage, supporting its potential as a rapidly deployable, shelf-stable platelet surrogate for use in emergency and austere settings.</div></div>","PeriodicalId":17326,"journal":{"name":"Journal of Thrombosis and Haemostasis","volume":"24 3","pages":"Pages 877-889"},"PeriodicalIF":5.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145505100","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimal adherence thresholds for oral anticoagulants in patients with atrial fibrillation using machine learning and population administrative data","authors":"Abdollah Safari ,&nbsp;Hamed Helisaz ,&nbsp;Mina Tadrous ,&nbsp;Marc W. Deyell ,&nbsp;Jason G. Andrade ,&nbsp;Shahrzad Salmasi ,&nbsp;Adenike Adelakun ,&nbsp;Kristian B. Filion ,&nbsp;Mary A. De Vera ,&nbsp;Peter Loewen","doi":"10.1016/j.jtha.2025.10.031","DOIUrl":"10.1016/j.jtha.2025.10.031","url":null,"abstract":"<div><h3>Background</h3><div>Adherence to oral anticoagulants (OACs) for atrial fibrillation (AF) stroke prevention is traditionally defined as taking 80% of doses as prescribed, but this threshold has not been clinically validated.</div></div><div><h3>Objectives</h3><div>We identified OAC adherence thresholds maximally associated with risk of clinical events in patients with AF and compared them with the conventional 80% threshold.</div></div><div><h3>Methods</h3><div>This was a cohort study using retrospective data of patients newly diagnosed with AF who were new OAC users from 1996 to 2019, based on population-based administrative data from British Columbia, Canada. We used Cox proportional hazards models with OAC proportion of days covered (PDC) as the main exposure captured during 90 days before outcome events or at the end of follow-up, then applied least absolute shrinkage and selection operator-estimated coefficient paths to identify candidate thresholds, and identified which were optimal in terms of outcomes.</div></div><div><h3>Results</h3><div>A total of 44 172 patients were included. For all outcomes, vitamin K antagonist (VKA) optimal thresholds were between PDC 0.85 and 0.95, and the direct OAC (DOAC) optimal threshold was PDC 0.9. Above vs below threshold outcome hazard reduction was greater for DOACs than for VKAs (DOACs: 19%-52% stroke risk reduction; VKAs: not significant for most outcomes). The PDC 0.8 threshold had inferior model fit and outcome effect compared with the optimal thresholds identified.</div></div><div><h3>Conclusion</h3><div>Optimal adherence thresholds are higher than conventionally assumed and, especially for DOACs, are strongly associated with clinical outcomes. These results provide a basis for considering updating the definition of OAC nonadherence to PDC &lt;0.9 for DOACs and to PDC &lt;0.9 or &lt;0.95 for VKAs, depending on the outcome.</div></div>","PeriodicalId":17326,"journal":{"name":"Journal of Thrombosis and Haemostasis","volume":"24 3","pages":"Pages 985-999"},"PeriodicalIF":5.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145505106","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparison of easy-to-use clinical prognostic models to identify low-risk normotensive patients with pulmonary embolism","authors":"Manuel Gloor ,&nbsp;Odile Stalder ,&nbsp;Tobias Tritschler ,&nbsp;Marie Méan ,&nbsp;Nicolas Rodondi ,&nbsp;Marc Righini ,&nbsp;Drahomir Aujesky","doi":"10.1016/j.jtha.2025.11.023","DOIUrl":"10.1016/j.jtha.2025.11.023","url":null,"abstract":"<div><h3>Background</h3><div>Lack of practicality is a common barrier to the use of clinical prognostic models for acute pulmonary embolism (PE).</div></div><div><h3>Objectives</h3><div>This study compared easy-to-use prognostic models with &lt;10 readily available variables to identify low-risk patients with PE who are candidates for home care.</div></div><div><h3>Methods</h3><div>Based on prospective cohort data from 677 patients with acute PE from 9 Swiss hospitals, we calculated 11 easy-to-use prognostic models (simplified version of Pulmonary Embolism Severity Index [sPESI], Geneva Prognostic Score, shock index [SI], BOVA, modified H-FABP, Syncope, Tachycardia [FAST], eStiMaTe, Platelets, Age, Troponin, Heart rate, Oxygenation, and Systolic blood pressure, Uresandi, Agterof, Huang, and European Society of Cardiology [ESC]-2019) and classified patients as low vs higher risk. The primary outcome was 30-day overall mortality. For each model, we examined test characteristics and the weighted clinical net benefit, ie, the percentage of avoided hospitalizations in low-risk patients minus the percentage of deceased low-risk patients, assuming that death is 100 times worse than a not avoided hospitalization.</div></div><div><h3>Results</h3><div>Overall, 21 of 677 patients (3.1%) died within 30 days. The proportion of patients classified as low risk varied widely from 18.9% (eStiMaTe) to 95.8% (SI). Mortality among low-risk patients ranged from 0% (sPESI, eStiMaTe, and ESC-2019) to 3.0% (modified FAST), with sensitivities varying from 100% (sPESI, eStiMaTe, and ESC-2019) to 9.5% (SI). Similarly, the negative likelihood ratios for mortality ranged from 0.06 (sPESI) to 0.95 (modified FAST). The weighted net benefit was highest for the sPESI (35.9%) and lowest for the SI (−187.6%).</div></div><div><h3>Conclusion</h3><div>The sPESI, the eStimMaTe score, and the ESC-2019 model most accurately identified patients with PE who are at low risk of mortality. The sPESI showed the highest clinical net benefit.</div></div>","PeriodicalId":17326,"journal":{"name":"Journal of Thrombosis and Haemostasis","volume":"24 3","pages":"Pages 1056-1066"},"PeriodicalIF":5.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145804882","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"“Effect of direct oral anticoagulants in cirrhosis: an in vitro study”: reply","authors":"Cindy Pereira Portela,&nbsp;Lucas A. Gautier,&nbsp;Alessandro Aliotta,&nbsp;Lorenzo Alberio,&nbsp;Debora Bertaggia Calderara,&nbsp;Maxime G. Zermatten","doi":"10.1016/j.jtha.2025.12.010","DOIUrl":"10.1016/j.jtha.2025.12.010","url":null,"abstract":"","PeriodicalId":17326,"journal":{"name":"Journal of Thrombosis and Haemostasis","volume":"24 3","pages":"Pages 1203-1206"},"PeriodicalIF":5.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147321889","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Large language models vs thrombosis experts: a comparative study on patient education and clinical decision-making in venous thromboembolism","authors":"Nikola Vladic ,&nbsp;Stephan Nopp ,&nbsp;Ingrid Pabinger ,&nbsp;Walter Ageno ,&nbsp;Jean M. Connors ,&nbsp;Sabine Eichinger ,&nbsp;Cihan Ay ,&nbsp;ClotGPT study group","doi":"10.1016/j.jtha.2025.09.004","DOIUrl":"10.1016/j.jtha.2025.09.004","url":null,"abstract":"<div><h3>Background</h3><div>Large language models (LLMs) have demonstrated remarkable capabilities in various medical fields, yet their performance in thrombosis and hemostasis, particularly in patient education and complex clinical decision making, is unexplored.</div></div><div><h3>Objectives</h3><div>We aimed to compare the quality of responses from LLMs vs thrombosis experts and assess clinician’s ability to distinguish between them.</div></div><div><h3>Methods</h3><div>Three experts on thrombosis and hemostasis and 3 LLMs (Le Chat Pixtral Large, DeepSeek-R1, and ChatGPT-4.5) answered 3 patient education and 3 clinical decision-making queries. Thirty-seven physicians rated responses for adequacy (1 = very poor; 10 = excellent) and estimated their origin (1 = certainly LLM; 10 = certainly human). Mean differences were assessed via <em>t</em>-tests, medians via Wilcoxon tests, and correlations via Spearman test. All <em>P</em> values were adjusted with Bonferroni correction.</div></div><div><h3>Results</h3><div>LLMs provided significantly better patient education responses than experts. Mean adequacy score differences were as follows: Le Chat Pixtral Large +1.6 (95% CI, 1.3-2.0; <em>P</em> &lt; .01), DeepSeek-R1 +1.7 (95% CI, 1.3-2.1; <em>P</em> &lt; .001), and ChatGPT-4.5 +1.9 (95% CI, 1.6-2.3; <em>P</em> &lt; .001). In clinical decision-making, DeepSeek-R1 outperformed experts (+1.4; 95% CI, 1.1-1.8; <em>P</em> &lt; .001), whereas Le Chat Pixtral Large (–0.3; 95% CI, –0.8-0.1; <em>P</em> = .96) and ChatGPT-4.5 (+0.5; 95% CI, 0.0-0.9; <em>P</em> = .18), performed comparably with experts. Evaluators could not distinguish between expert (median, 6.0; IQR, 3.0-8.0) and LLM-generated responses (median, 6.0; IQR, 4.0-8.0).</div></div><div><h3>Conclusion</h3><div>LLMs outperform experts in venous thromboembolism-related patient education and match or exceed them in clinical decision making, providing responses indistinguishable from experts. Although major barriers need to be addressed, LLMs have strong potential to support clinical management and patient education in the field of thrombosis and hemostasis.</div></div>","PeriodicalId":17326,"journal":{"name":"Journal of Thrombosis and Haemostasis","volume":"24 3","pages":"Pages 943-954"},"PeriodicalIF":5.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145182018","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}