Risks of anticoagulation discontinuation after first venous thromboembolism in inherited thrombophilia: a long-term cohort study Authors: Sang-A Kim, Sin Young Park, Ji Yun Lee, Junshik Hong, Sang-Hoon Na, Jeong-Ok Lee, Soo-Mee Bang Published: 2026 ABSTRACT AbstractBackgroundInherited thrombophilia confers high venous thromboembolism (VTE) recurrence risk, and guidelines generally recommend indefinite anticoagulation after first VTE. However, long-term data comparing outcomes between patients who continue versus discontinue therapy are scarce.ObjectivesTo compare recurrent VTE and major bleeding (MB) risks between patients who continued versus discontinued anticoagulation after first VTE in inherited thrombophilia.MethodsWe conducted a retrospective cohort study of 103 patients with protein S, protein C, or antithrombin deficiency who experienced first VTE between 1987 and 2023. Patients were grouped by anticoagulation status: 49 continued, 54 discontinued. The primary endpoint was recurrent VTE or MB.ResultsMedian age was 39 years (range, 11–81); 68% were male. Over median 6.0-year follow-up (interquartile range, 4.9–17.9 years), the composite endpoint occurred far more often in the discontinued versus continued group (66.7% vs 4.1%; HR, 20.23; 95% CI, 4.86–84.27; P < .001). Ten-year event-free survival was 95.1% versus 42.1% (P < .001), yielding a number needed to treat of 2. This difference was driven by VTE recurrence (63.0% vs 4.1%; adjusted HR, 19.85; P < .001). MB occurred only in the discontinued group (5.6% vs 0%). Anticoagulation discontinuation was the only independent predictor of recurrence. Time-dependent Cox and competing risk analyses confirmed these findings.ConclusionsIn patients with inherited thrombophilia, discontinuing anticoagulation markedly increased VTE recurrence without reducing bleeding. Our data support indefinite anticoagulation after first VTE in this population.Keywords: anticoagulants, cohort studies, secondary prevention, thrombophilia, venous thromboembolism Introduction Inherited thrombophilia refers to a group of genetic disorders associated with increased venous thromboembolism (VTE) risk [1,2]. The most common forms in Asian populations—protein S deficiency, protein C deficiency, and antithrombin deficiency—account for a substantial proportion of VTE cases in younger patients presenting without clear provocation. These deficiencies are regarded as high-risk inherited thrombophilias, with reported annual VTE recurrence rates of approximately 6% to 10% and cumulative 10-year recurrence exceeding 50% [3]. After the first VTE, patients with inherited thrombophilia experience significantly higher recurrence rates than those without these defects [3,4]. Current guidelines generally recommend indefinite anticoagulation following unprovoked or minimally provoked first VTE in this population, based on the concept that genetic abnormalities confer persistent thrombotic risk requiring ongoing therapy [[5], [6], [7]]. However, whether lifelong anticoagulation is both necessary and safe remains subject to clinical debate. Although prolonged anticoagulation reduces VTE recurrence, it carries increased major bleeding (MB) risk with potential for significant morbidity and mortality. Decisions about continuing or stopping anticoagulation require weighing competing risks of recurrent thrombosis against bleeding complications. Multiple studies have examined VTE recurrence after anticoagulation cessation in general VTE populations, but data specifically addressing inherited thrombophilia remain limited. Moreover, existing research has concentrated primarily on recurrence risk, with comparatively less focus on bleeding outcomes and net clinical benefit. The optimal anticoagulation duration in inherited thrombophilia thus remains unclear [8,9]. Real-world evidence comparing long-term outcomes between patients receiving indefinite anticoagulation and those discontinuing therapy is scarce, particularly concerning the balance of thrombotic versus bleeding risks across extended follow-up [1,10]. We performed this retrospective cohort study to evaluate long-term outcomes in inherited thrombophilia patients following their first VTE, comparing recurrent VTE and MB risks between those receiving indefinite anticoagulation and those discontinuing therapy. Our primary aim was to determine whether anticoagulation cessation affects the composite outcome of recurrent VTE or MB, providing evidence to guide treatment duration decisions in this high-risk group. Methods We conducted a retrospective cohort study at the Seoul National University Bundang Hospital and Seoul National University Hospital, reviewing records of patients with a first episode of VTE and diagnosed with inherited thrombophilia between January 1987 and December 2023. The study was approved by the institutional review board of each hospital (IRB No.: B-2502-954-107 and H-2502-174-1620, respectively), and informed consent was waived due to the retrospective nature. Eligible patients had objectively confirmed VTE and received anticoagulation therapy. Patients were included if adequate medical records for the index VTE episode were available, regardless of whether the initial event was treated at our institutions. Inherited thrombophilia (protein S, protein C, or antithrombin deficiency) was confirmed through functional assays and, where available, genetic testing. Testing was performed at least 3 months after the acute VTE event, when the patient was free of conditions that could interfere with assay results (eg, ongoing anticoagulation, acute thrombosis, pregnancy, inflammatory states). Provoking factors at the time of index VTE were categorized according to the International Society on Thrombosis and Haemostasis (ISTH) Scientific and Standardization Committee 2016 guidance as major transient (major surgery, trauma, or hospital immobilization ≥ 3 days), minor transient (combined oral contraceptives, puerperium, or long-haul air travel > 8 hours), or absent (unprovoked) [11]. Exclusion criteria included incomplete medical records, unknown anticoagulation status, or incomplete anticoagulation timing data. Patients were classified based on anticoagulation status at the last follow-up. The continued anticoagulation group consisted of patients continuing anticoagulation without a predefined stopping point. The discontinued group included patients whose anticoagulation was stopped after a defined course of therapy. The decision to continue or discontinue anticoagulation was made by the treating physician in consultation with the patient, considering individual patient factors, bleeding risk, and patient preference. Outcomes The primary endpoint was a composite of recurrent VTE or MB, assessed from the date of index VTE to the date of first event, death, or last follow-up. Recurrent VTE was defined as a new objectively confirmed symptomatic deep vein thrombosis (DVT), pulmonary embolism, or other venous thrombotic event occurring after completion of initial anticoagulation for the index event. Diagnosis required confirmation by compression ultrasonography, computed tomography, and/or magnetic resonance imaging. MB was defined according to the ISTH criteria as follows: (1) fatal bleeding; (2) bleeding in a critical site (intracranial, intraspinal, intraocular, pericardial, intra-articular, intramuscular with compartment syndrome, or retroperitoneal); (3) bleeding causing a fall in hemoglobin level of 2 g/dL or more; or (4) bleeding leading to transfusion of 2 or more units of whole blood or packed red blood cells. Secondary endpoints included recurrent VTE, MB, and clinically relevant nonmajor bleeding (CRNMB) [12,13]. CRNMB was defined according to ISTH criteria as overt bleeding that did not meet criteria for MB but was associated with medical intervention, unscheduled contact with a physician, temporary cessation of anticoagulation, or associated with discomfort or impairment of activities of daily life. Any bleeding includes both MB and CRNMB. Statistical analysis Continuous variables are presented as median with interquartile range or mean with standard deviation, as appropriate. Categorical variables are expressed as frequencies and percentages. Baseline characteristics were compared between groups using the Student’s t-test or Mann–Whitney U-test for continuous variables and the chi-square test or Fisher’s exact test for categorical variables. The primary analysis evaluated time to the composite endpoint using Kaplan–Meier survival analysis, with comparison between groups using the log-rank test. Cox proportional hazards regression was used to estimate hazard ratios (HRs) with 95% CI for the association between anticoagulation discontinuation and outcomes. Multivariable Cox models were adjusted for age, sex, thrombophilia type (protein S, protein C, or antithrombin deficiency), provoking factor category (major transient, minor transient, or unprovoked), and atypical site thrombosis (cerebral or splanchnic vein thrombosis). For secondary endpoints, similar survival analyses were performed. For MB outcomes where one group had zero events, Fisher's exact test was used instead of Cox regression. Sensitivity analyses included: (1) addition of all-cause mortality to the composite endpoint (2) time-dependent Cox regression accounting for anticoagulation exposure as a time-varying covariate, and (3) reclassification of patients lost to follow-up to the continued anticoagulation group (censored at their last recorded follow-up date). The proportional hazards assumption was verified for all Cox models. All statistical tests were 2-sided, and a P value < .05 was considered statistically significant. All analyses were performed using R version 4.4.2 (R Foundation for Statistical Computing, Vienna, Austria). Results Baseline characteristics and anticoagulation A total of 103 patients with inherited thrombophilia who experienced a first VTE were included in the analysis. Of the 103 patients, 49 (47.6%) continued indefinite anticoagulation (continued group) and 54 (52.4%) discontinued anticoagulation after a defined period (discontinued group). Baseline characteristics are summarized in Table 1. The median age at index VTE was 39 years (range, 11–81 years), and 70 patients (68.0%) were male. Protein S deficiency was the most common subtype, affecting 55 patients (53.4%), followed by protein C deficiency in 29 (28.2%) and antithrombin deficiency in 19 (18.4%). The distribution of natural anticoagulant activity levels within each deficiency subgroup is shown in Supplementary Figure 1. Inherited thrombophilia was diagnosed based on either genetic confirmation (n = 30, 29.1%) or persistently reduced levels of natural anticoagulants documented on at least 2 consecutive measurements after exclusion of secondary causes (n = 73, 70.9%). Among genetically confirmed cases, 13 had protein S deficiency, 13 had protein C deficiency, and 4 had antithrombin deficiency. Table 1Baseline patient characteristicsaVariableTotal (n = 103)Continued anticoagulation (n = 49)Discontinued anticoagulation (n = 54)P valueMale70 (68.0)33 (67.3)37 (68.5)1.000Asian103 (100.0)49 (100.0)54 (100.0)-Age (y) (median, range)39 (11–81)41 (28–51)36 (26–47).376Diagnosis.087 Protein S deficiency55 (53.4)31 (63.3)24 (44.4) Protein C deficiency29 (28.2)9 (18.4)20 (37.0) Antithrombin deficiency19 (18.4)9 (18.4)10 (18.5)Secondary thrombophilic conditionb.426 Major transient risk factor Trauma4 (3.9)2 (4.1)2 (3.7) Immobilization5 (4.9)3 (6.1)2 (3.7) Major surgery5 (4.9)3 (6.1)2 (3.7) Solid cancer (curatively treated)3 (2.9)1 (2.0)2 (3.7) Minor transient risk factor Puerperium1 (1.0)0 (0.0)1 (1.9) Oral contraceptive3 (2.9)1 (2.0)2 (3.7) Air travel >8 h2 (1.9)1 (2.0)1 (1.9)Initial VTE diagnosis.331 PE20 (19.4)7 (14.3)13 (24.1) DVT36 (35.0)19 (38.8)17 (31.5) PE and DVT28 (27.2)14 (28.6)14 (25.9) CSVT3 (2.9)0 (0.0)3 (5.6) SVT15 (14.6)8 (16.3)7 (13.0) Mixed/Other1 (1.0)1 (2.0)0 (0.0)Initial anticoagulant for index VTE<.001 Warfarin46 (44.7)13 (26.5)33 (61.1) LMWH2 (1.9)02 (3.7) DOAC55 (53.4)36 (73.5)19 (35.2)Intervention.622 Radiologic mechanical thrombectomy5 (4.9)3 (6.1)2 (3.7) Surgical thrombectomy3 (2.9)2 (4.1)1 (1.9) Thrombolysis16 (15.5)6 (12.2)10 (18.5)aN represents the number of patients with each characteristic. Data are presented as n(%) unless otherwise specified.bPatients may have had more than one risk factor. CSVT, cerebral sinus venous thrombosis; DOAC, direct oral anticoagulant; DVT, deep vein thrombosis; LMWH, low-molecular-weight heparin; PE, pulmonary embolism; SVT, splanchnic vein thrombosis; VTE, venous thromboembolism. Baseline patient characteristicsa N represents the number of patients with each characteristic. Data are presented as n(%) unless otherwise specified. Patients may have had more than one risk factor. CSVT, cerebral sinus venous thrombosis; DOAC, direct oral anticoagulant; DVT, deep vein thrombosis; LMWH, low-molecular-weight heparin; PE, pulmonary embolism; SVT, splanchnic vein thrombosis; VTE, venous thromboembolism. The index VTE involved isolated DVT in 36 patients (35.0%), pulmonary embolism with or without DVT in 48 (46.6%), and atypical sites in 19 patients (18.4%) including cerebral venous sinus thrombosis (n = 3, 2.9%), splanchnic vein thrombosis (n = 15, 14.6%), and combined DVT with splanchnic thrombosis (n = 1, 1.0%). Sixteen patients (15.5%) had at least 1 provoking factor at the time of index VTE, which were categorized according to the ISTH SSC 2016 framework as major transient (n = 11; including major surgery, trauma, hospital immobilization ≥ 3 days, or limited-stage cancer treated with curative intent) or minor transient (n = 5; including combined oral contraceptives, puerperium, or long-haul air travel > 8 hours). All 3 patients with cancer-associated index VTE had limited-stage disease treated with curative intent and were in complete remission at the time of analysis; accordingly, they were classified within the major transient category. The anticoagulant initially prescribed for treatment of the index VTE was direct oral anticoagulant (DOAC) in 55 patients (53.4%), warfarin in 46 (44.7%), and low-molecular weight heparin in 2 (1.9%), with DOAC use significantly higher in the continued group (73.5% vs 35.2%, P < .001). Among DOAC-treated patients, rivaroxaban was used in 25, dabigatran in 20, edoxaban in 6, and apixaban in 4. During extended treatment, all patients receiving rivaroxaban or apixaban were transitioned to the labeled reduced dose (10 mg once daily and 2.5 mg twice daily, respectively); 3 of 6 patients receiving edoxaban received 30 mg once daily, and all patients receiving dabigatran remained on full dose. The median duration of anticoagulation in the discontinued group was 1.0 year (range, 0.1–19.2 years). Among patients who discontinued anticoagulation, the most common reasons were physician’s discretion (n = 22, 40.7%), loss to follow-up (n = 7, 13.0%), self-discontinuation (n = 5, 9.3%), and bleeding complications (n = 4, 7.4%; 2 MB, 2 CRNMB). Primary endpoint Over the median follow-up period of 6.0 years (interquartile range, 4.9–17.9 years), the median follow-up time of each group was 7.7 years (range, 0.1–47.7 years) in continued group and 5.4 years (range, 0.4–21.7 years) in discontinued group, with no significant difference between groups (P = .056). The composite primary endpoint of recurrent VTE or MB occurred in 2 of 49 patients (4.1%) in the continued group compared with 36 of 54 patients (66.7%) in the discontinued group. The 10-year event-free survival rates were 95.1% (95% CI, 88.6%–100.0%) in continued group versus 42.1% (95% CI, 30.2%–58.8%) in the discontinued group (P < .001) (Figure A). Discontinuation of anticoagulation was associated with about 20-fold increased risk of the composite endpoint (HR, 20.23; 95% CI, 4.86–84.27). Kaplan–Meier analysis of anticoagulation discontinuation. (A) Event-free survival for the composite endpoint of recurrent venous thromboembolism or major bleeding. (B) Venous thromboembolism recurrence-free survival. Solid line, continued anticoagulation (n = 49); dashed line, discontinued anticoagulation (n = 54). Tick marks indicate censored observations. Numbers at risk are shown at 5-year intervals. Secondary endpoints Recurrent VTE occurred in 2 of 49 patients (4.1%) in the continued group and 34 of 54 patients (63.0%) in the discontinued group (HR, 17.82; 95% CI, 4.28–74.22; P < .001). The 10-year recurrence-free survival rates were 95.1% (95% CI, 88.6%–100.0%) versus 44.2% (95% CI, 32.1%–60.8%), respectively (Figure B). Among patients in the discontinued group who experienced recurrence (n = 34), the median time to recurrence was 3.6 years (range, 0.4–13.8 years). Recurrent VTE events occurred continuously over the follow-up period, with events observed over 10 years after the index VTE. There was no significant relationship between anticoagulation duration and VTE recurrence (P = .070). All recurrent events were managed with anticoagulation resumption; none required intensive care unit admission or resulted in VTE-related death. MB occurred in 0 of 49 patients (0%) in the continued group and 3 of 54 patients (5.6%) in the discontinued group (P = 0.244). The 3 MB events in the discontinued group included 2 gastrointestinal bleeding and 1 intracranial bleeding events. No fatal bleeding events occurred in either group. Clinically relevant nonmajor bleeding occurred in 2 of 49 patients (4.1%) in the continued group and 3 of 54 patients (5.6%) in the discontinued group (P = 1.000). In total, any bleeding occurred in 2 of 49 patients (4.1%) in the continued group and 6 of 54 patients (11.1%) in the discontinued group (P = 0.274). Overall, VTE recurrence events (n = 36) far outnumbered MB events (n = 3). In univariable Cox regression analysis, anticoagulation discontinuation was strongly associated with VTE recurrence. Neither major transient nor minor transient risk factors reached statistical significance (major transient HR, 2.21; 95% CI, 0.91–5.35; P = .078; minor transient HR, 2.82; 95% CI, 0.85–9.30; P = .089), and age, sex, thrombophilia type, and atypical site thrombosis were not significantly associated with recurrence (Table 2). In multivariable analysis, discontinuation of anticoagulation remained the only independent predictor of VTE recurrence (adjusted HR, 19.85; 95% CI, 4.68–84.23; P < .001). Table 2Univariable and multivariable analyses for recurrent venous thromboembolism.FactorsNUnivariable HR (95% CI)PMultivariable HR (95% CI)PAge-0.99 (0.97–1.01).1740.98 (0.96–1.01).149Female330.67 (0.32–1.43).3010.61 (0.27–1.34).213Thrombophilia type.062.135 Protein S deficiency551 (reference)1 (reference) Protein C deficiency291.36 (0.65–2.83).4100.82 (0.39–1.76).614 Antithrombin deficiency190.99 (0.41–2.41).9850.85 (0.34–2.14).729Provoking factor Unprovoked871 (reference)(reference) Major transient112.21 (0.91–5.35).0781.74 (0.67–4.54).254 Minor transient52.82 (0.85–9.30).0892.82 (0.66–11.96).163Atypical site thrombosis190.48 (0.17–1.36).1690.51 (0.17–1.55).237Anticoagulation discontinuation5417.82 (4.28–74.22)<.00119.85 (4.68–84.23)<.001HR, hazard ratio. Univariable and multivariable analyses for recurrent venous thromboembolism. HR, hazard ratio. Sensitivity analysis In sensitivity analysis including all-cause mortality as a competing outcome, anticoagulation discontinuation remained significantly associated with the composite endpoint (HR, 11.51; 95% CI, 4.01–33.03; P < .001). Time-dependent Cox regression analysis, accounting for the duration of anticoagulation therapy as a time-varying covariate, also confirmed the markedly increased risk of recurrent VTE during periods off anticoagulation (HR, 20.90; 95% CI, 6.60–66.20; P < .001) (Supplementary Figure 2). In an additional sensitivity analysis, reclassification of the 7 patients lost to follow-up to the continued anticoagulation group (with censoring at their last recorded follow-up) yielded a consistent association between anticoagulation discontinuation and the composite endpoint (HR, 9.42; 95% CI, 3.90–22.74; P < .001). Discussion In this retrospective cohort study of 103 patients with inherited thrombophilia following first VTE, anticoagulation discontinuation was associated with substantially increased risk of adverse outcomes. The composite endpoint of recurrent VTE or MB occurred in 66.7% of patients who discontinued anticoagulation versus 4.1% who continued indefinitely (HR, 20.23; 95% CI, 4.86–84.27; P < .001). This difference stemmed primarily from VTE recurrence (63.0% vs 4.1%) rather than reduced bleeding—in fact, MB rates were higher in the discontinuation group (5.6% vs 0%). These findings offer important real-world evidence favoring indefinite anticoagulation in this high-risk population. The high recurrence rate in our discontinued group (63.0% over median 5.4-year follow-up) matches previous studies reporting 20% to 50% recurrence in inherited thrombophilia, though ours is higher [3,[14], [15], [16]]. This probably reflects our longer overall follow-up (median 6.0 years, with recurrences up to 13.8 years) compared with typical 2- to 5-year studies, showing that recurrence risk persists long-term. Our finding that indefinite anticoagulation reduces recurrence to 4.1% fits with the biological concept that genetic thrombophilia creates permanent thrombotic risk [17]. Although some studies have reported higher recurrence rates in antithrombin deficiency compared with protein S or C deficiency, we found no significant difference in recurrence risk among thrombophilia subtypes (P = 0.135). This may reflect our relatively small sample size, and further studies with larger cohorts are needed to clarify whether recurrence risk differs by deficiency type. The absence of increased bleeding with indefinite anticoagulation (0% vs 5.6% MB) challenges the assumption that prolonged anticoagulation must increase bleeding risk. Several explanations are possible: patients selected for indefinite therapy may have inherently lower bleeding risk, closer monitoring may prevent events, or bleeding may have occurred during anticoagulation before discontinuation. Additionally, the higher proportion of DOAC use in the continued group (73.5% vs 35.2%, P < .001) may have contributed, as the convenience of fixed dosing and no routine monitoring may facilitate long-term adherence. Our composite endpoint incorporating both thrombotic and bleeding outcomes allows complete assessment of net clinical benefit [18]. The 10-year event-free survival (95.1% vs 42.1%) translates to a number needed to treat of approximately 2, a favorable benefit-risk balance. Although no VTE-related deaths occurred and recurrent events were manageable (none requiring intensive care unit admission), recurrent VTE brings significant morbidity and healthcare costs, justifying a prevention approach. Our findings support current guidelines recommending indefinite anticoagulation in patients with these high-risk inherited thrombophilia after the first VTE [[5], [6], [7]]. Discontinuation should occur only in exceptional circumstances following careful discussion of recurrence risk. For patients stopping therapy (whether by physician decision or patient preference), close surveillance and a low threshold for restarting anticoagulation are essential. This study has several strengths. The relatively long follow-up (median 6.0 years, events up to 13.8 years) provides solid long-term outcome data from real-world practice. We assessed both thrombotic and bleeding outcomes comprehensively, evaluating net clinical benefit rather than recurrence alone. All patients had active clinical follow-up with complete event capture through electronic medical records. Findings remained consistent across multiple sensitivity analyses—time-dependent Cox regression (HR, 20.90; P < .001) and all-cause mortality inclusion (HR, 11.51; P < .001)—confirming result reliability. Our study was restricted to deficiencies of the natural anticoagulants—protein C, protein S, and antithrombin—and did not include patients with factor (F)V Leiden (FV) or the prothrombin G20210A mutation, both of which are well-recognized inherited thrombophilias associated with substantially increased recurrence risk, particularly in homozygous forms or in combined heterozygous FV and heterozygous prothrombin G20210A. However, both variants are exceedingly rare in East Asian populations, as demonstrated in a recent systematic review and meta-analysis of East Asian VTE cohorts and in a Korean thrombophilia testing cohort [19,20]. In this regard, our cohort may offer complementary long-term outcome data on a group of inherited thrombophilias that are comparatively less common in Western populations. An unexpected finding in our initial analysis was that the presence of a provoking factor at index VTE appeared to be associated with higher recurrence risk in univariable analysis. On reclassification of provoking factors according to the ISTH SSC 2016 guidance, neither major transient nor minor transient risk factors were significantly associated with recurrence, suggesting that the apparent univariable association observed with the combined “provoked VTE” variable reflected pooling of a heterogeneous group of factors rather than a true biological effect [11]. Notably, the duration of anticoagulation in the discontinued group was relatively short in both the provoked and unprovoked subgroups (median approximately 1 year), suggesting that in patients with inherited thrombophilia, this duration is insufficient to confer lasting protection against recurrence, regardless of whether a transient provoking factor was identified at the index event. These findings support the interpretation that in this high-risk population, the underlying thrombogenic diathesis persists even after resolution of any identifiable provoking factor, and categorization as “provoked” should not, by itself, be used to justify anticoagulation cessation. Important limitations merit consideration. The retrospective, observational design introduces potential selection bias, as anticoagulation decisions reflected clinical judgment rather than randomization. However, this approach is consistent with major studies in this field [14,16,21], where ethical considerations preclude randomized discontinuation trials in high-risk populations. The large effect size makes unmeasured confounding unlikely to fully explain our findings. Some patients discontinued due to bleeding (n = 4, 7.4%), but the overwhelming predominance of VTE recurrence (34 events) over MB (3 events) indicates this did not materially affect our conclusions. Retrospective bleeding ascertainment may underestimate minor events, though MB requiring medical attention is unlikely to be missed. Additionally, the type of anticoagulation differed between groups, with higher DOAC use in the continued group, which may have influenced outcomes. Finally, the small number of patients with identifiable provoking factors at index VTE, particularly within the minor transient category (n = 5), resulted in wide CIs for these subgroup estimates; the corresponding hazard ratios should therefore be interpreted with caution. In conclusion, discontinuation of anticoagulation in inherited thrombophilia was associated with substantially increased VTE recurrence without reduced bleeding, resulting in worse net clinical outcomes. With a number needed to treat of only 2, indefinite anticoagulation demonstrates a compelling benefit-risk profile in patients with these high-risk inherited thrombophilias. These findings support indefinite anticoagulation as the default approach in patients with inherited thrombophilia following first VTE, with discontinuation reserved for exceptional circumstances and accompanied by close monitoring. Acknowledgments Author contributions S.A. Kim collected data, performed statistical analysis, and wrote the manuscript. S.Y. Park collected data. J.Y. Lee, J. Hong, S.H. Na, and J.O. Lee provided patient data. S.M. Bang designed the study, collected data, and supervised the manuscript. All authors critically reviewed and approved the final version. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Relationship Disclosure The authors declare that they have no conflicts of interest. Declaration of generative AI in scientific writing The manuscript was primarily written by S.A. Kim. Generative AI (Claude, Anthropic) was used for English-language editing. Statistical analyses were primarily performed by S.A. Kim, with AI assistance used only for troubleshooting R code. The authors take full responsibility for the content of this publication. Footnotes Handling Editor: Dr Choo Fang The online version contains supplementary material available at https://doi.org/10.1016/j.rpth.2026.106788. Supplementary material Supplementary Figure S1 and S2 References Associated Data Supplementary Materials Supplementary Figure S1 and S2