Summary
A medRxiv preprint used proteome-wide Mendelian randomization to identify 11 plasma proteins associated with early-onset stroke, defined as stroke occurring before age 60. The strongest signals involved coagulation and thromboinflammatory pathways, with Factor XI highlighted as a potential therapeutic target.
A large-scale genetic analysis has identified 11 plasma proteins with likely causal associations with early-onset stroke, which the study defines as stroke occurring before age 60. The preprint, posted on medRxiv on September 15, 2026, found that several of the strongest signals were connected to blood coagulation and inflammation-related processes, and highlighted Factor XI as a leading target for future prevention research.
The researchers tested 1,041 proteins using two-sample Mendelian randomization. Their analysis drew on protein-related genetic data from 54,219 UK Biobank participants and stroke association data covering up to 11,114 early-onset stroke cases and 435,540 controls.
Eleven proteins showed significant associations
After correction for multiple testing, genetically predicted levels of 11 proteins were significantly associated with early-onset stroke. Higher predicted levels of proteins encoded by ABO, F11, F7, MLN, BTD, MEP1B and VAMP8 were associated with increased risk. The reported odds ratios ranged from approximately 1.05 to 1.32 for each standard-deviation increase in genetically predicted protein level.
Higher predicted levels of four other proteins—FN1, GRK5, EHBP1 and FCN2—were associated with lower risk. The study also identified eight additional proteins whose associations varied by stroke subtype. Those subtypes were classified using the TOAST system, which groups ischemic strokes according to their likely underlying mechanism.
The signals involving ABO, F11 and F7 were particularly consistent across the analyses. The researchers also found overlap between the protein associations and genetic risks related to late-onset stroke, venous thromboembolism and cardiometabolic traits. This pattern supports the possibility that some biological mechanisms are shared across different vascular diseases and age groups.
How the genetic analysis works
Mendelian randomization uses naturally occurring genetic variants as instruments for an exposure—in this case, the level of a circulating protein. Because genetic variants are assigned before disease develops, the method can help assess whether a protein may have a causal relationship with an outcome rather than merely appearing alongside it in observational data.
For this analysis, the researchers selected variants within 200 kilobases of protein-coding genes that were associated with protein levels at a threshold of p ≤1×10−5. The variants were then clumped to reduce linkage disequilibrium, a situation in which nearby genetic variants are inherited together. Their effects were estimated using inverse-variance weighted Mendelian randomization, alongside sensitivity analyses intended to examine the robustness of the results.
This design makes the findings mechanistic and target-prioritisation evidence rather than results from a clinical trial or a test of a finished diagnostic panel. The associations describe genetically predicted protein levels, so further clinical research is needed to determine how well these proteins measure individual stroke risk in practice.
Factor XI emerges as a treatment research target
The study’s drug-target mapping placed Factor XI, encoded by F11, among the leading candidates for therapeutic investigation. Factor XI participates in the coagulation cascade, the chain of reactions that helps form blood clots. The researchers noted that existing Factor XI inhibitors, including abelacimab and milvexian, provide examples of compounds relevant to this target area.
The result does not test either drug in this study. Instead, it connects a genetically predicted coagulation-related protein signal with a class of potential interventions that can be examined in future prevention studies.
Overall, the preprint points to coagulation, platelet activation, extracellular-matrix remodelling and immune pathways as contributors to early-onset stroke biology. Its findings could help prioritise proteins for laboratory validation, risk-prediction research and clinical studies focused on preventing stroke at younger ages. As a medRxiv preprint based on summary-level genetic data, the work represents an early research-stage assessment rather than a clinical recommendation.