Phase-specific prognostic factors for recurrent venous thromboembolism during and after anticoagulation: a prospective cohort study from a Chinese tertiary thrombosis center

Long-term recurrence patterns of venous thromboembolism (VTE) in Chinese patients remain insufficiently characterized, particularly across treatment phases. In this prospective cohort study, 581 consecutive adults with a first objectively confirmed symptomatic VTE were longitudinally followed at a Chinese tertiary thrombosis and haemostasis center. Using analysis-specific time origins, recurrence risk was evaluated during anticoagulation, after anticoagulation discontinuation, and from a 3-month landmark analysis. Fine–Gray models accounted for competing death during anticoagulation, and Kaplan–Meier estimates and Cox models were used after anticoagulation discontinuation and in the landmark analysis. The cohort was characterized by a young age at VTE onset (median age, 36 years), relatively high proportions of unusual-site VTE (26.0%), unprovoked VTE (68.7%), and inherited thrombophilia (39.4%). Median observed anticoagulation duration was 12 months (IQR, 6–25). During anticoagulation, 34 recurrences occurred over 902.7 patient-years (3.8/100 patient-years), and older age was associated with recurrence (adjusted sHR per decade, 1.358). After anticoagulation discontinuation, 118 recurrences occurred over 798.0 patient-years (14.8/100 patient-years), with cumulative risk of 21.3%, 36.4%, and 49.4% at 12 months, 36 months, and 84 months, respectively. Older age at VTE onset (aHR per decade, 1.330), inherited thrombophilia (aHR, 1.837), and provoking status (global P = 0.016) were associated with recurrence. In the 3-month landmark analysis, continuation of anticoagulation beyond 3 months was associated with a lower recurrence risk than discontinuation (aHR, 0.537). In conclusion, recurrent VTE showed distinct phase-specific patterns, with a markedly higher recurrence rate after anticoagulation discontinuation. Older age at VTE onset was consistently associated with recurrence across phases, whereas inherited thrombophilia and provoking status were mainly associated with recurrence after anticoagulation discontinuation. Continuation of anticoagulation beyond 3 months was associated with a lower recurrence risk, but this observational association should not be interpreted as causal.