Toward precision antithrombotic management

In their lead article "Precision cardiovascular medicine: shifting the innovation paradigm", Aikawa et al. address a timely and highly relevant topic (Aikawa et al, 2025). The article delivers several key messages. On one hand, it summarizes recent technological advancesmost notably, the revolution in artificial intelligence (AI)-that now enable large-scale characterization of heterogeneous diseases and patient populations. On the other hand, it paves the way for concerted efforts that could ultimately translate these technological advances into personalized cardiovascular medicine.A crucial next step is to foster collaboration. Given the complexity and specialization of the required technical skills, successful translation to clinical practice demands close cooperation between all stakeholders, including physicians and patients, government bodies, and industry partners. In these challenging times, with growing restrictions on public funding for innovation, such collaborations-both national and international-are more critical than ever to ensure efficient progress. This discussion is also timely because cardiovascular medicine has lagged behind oncology in adopting precision approaches. While this comparison may seem unfair-oncologists, after all, deal with highly heterogeneous malignancies-the molecular revolution has enabled deep genomic and phenotypic profiling in cancer care, leading to advances in personalized oncology. Cardiovascular disease (CVD) is even more complex than cancer, owing to its multifactorial and polygenic nature, and yet a similar precision medicine approach is becoming increasingly feasible thanks to breakthroughs in diagnostics, imaging, bioinformatics, and AI.Even without the full application of the comprehensive methodologies described by Aikawa et al., the field of cardiovascular medicine is already trending toward personalization. First, there is growing awareness of sex-based biological differences in disease presentation, pathophysiology (e.g., plaque erosion vs. rupture), and response to therapy (Manzo-Silberman et al. 2024). Although this has not yet translated into sex-specific guidelines, research in this area-also including DNA-guided medication prescription-is expanding.Second, there is increasing advocacy for a holistic physician-patient relationship, recognizing the role of patient knowledge, preferences, and motivation in therapy adherence. Shared decision-making is now embedded in contemporary guidelines-such as those for atrial fibrillation (AF)-which include sections on patient context, emphasizing comprehensive care plans that consider side effects, adherence, and long-term goals (van Gelder et al. 2024).The next frontier is truly individualized pharmacotherapy. Progress here has been slow, as exemplified by the use of antithrombotic medication.Myocardial infarction (MI) and ischemic stroke account for 85% of all deaths due to CVD worldwide (WHO website). Thrombosis is the central event in these conditions, as well as in venous thromboembolism (VTE). Millions globally-men and women alike-are prescribed antithrombotic agents, and approximately one in three individuals will develop AF, often requiring anticoagulation (Proietti & Yip 2020).Before 1970, the antithrombotic arsenal was limited to aspirin, heparin, and warfarin. While warfarin and heparin therapy were individualized using laboratory parameters (e.g., activated partial thromboplastin time and international normalized ratio; INR), treatment success relied heavily on maintaining patients within a narrow therapeutic range-typically an INR of 2-3.Despite being a form of early personalized medicine, the practical difficulties of maintaining a stable INR made this strategy vulnerable. The advent of direct oral anticoagulants (DOACs), which offer fixed dosing without laboratory monitoring, marked a shift toward convenience, but arguably at the expense of individualized care (ten Cate 2025).DOACs have more predictable pharmacokinetics than vitamin K antagonists (VKA), obviating routine laboratory monitoring. However, a recent randomized trial in frail elderly patients with AF showed higher bleeding rates with DOACs compared with INR-guided VKA therapy, challenging assumptions about their safety (Joosten et al. 2024). One relevant factor for this unexpected enhanced bleeding risk, among several possible factors, may be the variation in DOAC plasma levels seen in such elderly populations. Substantial interindividual variation in DOAC plasma levels-linked to both thrombotic and bleeding risks-has been well documented, yet remains largely unmeasured in clinical practice. This represents a "known unknown" that compromises safety in certain populations (ten Cate 2025). One potential improvement would be to reintroduce lab-guided dosing based on direct measurement of DOAC plasma levels rather than on INR. This could offer a path toward repersonalizing anticoagulant management without reverting to the burdens of VKA monitoring.To truly achieve precision medicine in thrombosis, we must go further. As Aikawa et al.argue (Aikawa et al. 2025), comprehensive patient phenotyping could enable more accurate risk stratification, in this case for both thrombosis and bleeding. For instance, in unprovoked VTE, the risk of recurrence after stopping anticoagulation is approximately 20% at 2 years, while the annual risk of major bleeding on continued therapy is approximately 2-3%. Current risk prediction models are inadequate at the individual level. Laboratory biomarkers, including thrombin generation assays and viscoelastic tests such as thromboelastography, may help close this gap, but supporting clinical evidence is still scarce (Shaw et al. 2025).Ten Cate: Precision antithrombotic management recurrence, cancer, or arterial events (Iding et al. 2023). Incorporating such tools into individualized risk scores could greatly improve treatment decisions.With emerging anticoagulants such as factor XI(a) (FXIa) inhibitors, identifying the right patient becomes even more crucial. Unlike DOACs and VKA, which act on central pathways, FXI is a peripheral clotting factor. Its role in thrombosis may be conditional and bypassable via alternative pathways, such as direct factor IX activation by kallikrein. This means FXI inhibition might only be effective in certain contexts or patient subgroups. Biomarkers detecting activated FXI in blood may theoretically help to select patients for such specific anticoagulant drugs, thus clarifying patient heterogeneity to limit use of selective antithrombotic agents to those that may benefit most. A similar logic applies to novel antiplatelet agents targeting glycoprotein VI (GPVI) or C-type lectin-like receptor 2 (CLEC-2). These may be beneficial in subsets of patients with acute ischemic stroke but not in all stroke patients and (point-of-care) assays that detect specific platelet epitopes, such as GPVI or circulating biomarkers, may help to identify those at greatest benefit.The heterogeneity of diseases such as AF and VTE demands better patient profiling-not just to identify relevant thrombotic pathways, but also to understand the dynamic and multifactorial nature of these conditions. Thromboinflammatory mechanisms are central to conditions such as MI and stroke, but targeted drug treatment is still in its early days. History offers cautionary lessons. Three decades ago, promising therapies (activated protein C, antithrombin, tissue pathway factor inhibitor) failed in clinical sepsis trials despite convincing efficacy data in primate sepsis models. A major reason was the heterogeneity of sepsis in humans-something not reflected in standardized animal models. Other issues relate to limitations in drug dosing given the bleeding potential of these agents in patients with sepsis associated coagulopathy. Lacking tools to stratify or subgroup patients, these trials were doomed to miss potential responders.Despite vastly improved technologies, we still risk repeating similar mistakes. A recent example is the OCEANIC-AF trial of the FXIa inhibitor asundexian, which failed to demonstrate superiority over apixaban in patients with AF (Piccini et al. 2025). Whether this reflects underdosing, biochemical escape pathways, or patient heterogeneity is unclear. But it reinforces the need for deep phenotyping: AF, though generally thrombogenic, encompasses a wide spectrum of risk profiles. While oral anticoagulants outperform antiplatelet agents in preventing stroke, the pathophysiology and specifically the coagulation pathways that are engaged (including FXI activation or not) likely differ among patients. Grouping all AF patients under the same anticoagulation strategy-especially with novel agents acting on noncentral pathways-may miss therapeutic windows in key subgroups (Rocca & Ten Cate 2025).As Aikawa and colleagues state: "Heterogeneity in the complex pathobiology and presentations of cardiovascular diseases (CVDs) limits the effectiveness of conventional "one size fits all" therapies for some of the world's leading causes of death" (Aikawa et al. 2025).In sum, precision medicine in thrombosis management holds promise for improving efficacy and safety through better risk stratification, therapy selection, and monitoring. Nextgeneration antithrombotics targeting non-essential pathways such as GPVI, FXI, FXII, and kallikrein may offer reduced bleeding risks-but only if matched to the right patients.Because these factors are not universally active, individualized profiling becomes indispensable.