目的：為探討口服抗凝血藥rivaroxaban 用於膝關節(total knee replacement, TKR) 或髖關節置換 (total hip replacement, THR) 術後預防靜脈栓塞症 (venous thromboembolism, VTE) 之療效與安全性，並與warfarin 及未用藥組做比較。
方法：採回溯性世代研究，收集2011 年至2014 年接受關節置換術病人，並追蹤6 個月。排除年齡小於18 歲、住院前3 個月內發生血栓栓塞症、術前1 年有心房纖維顫動病史或曾接受心臟瓣膜置換手術，及術後warfarin 與rivaroxaban 互換的病人等。病人資料以Kaplan-Meier analysis 及log-rank test 進行分析。
結果：共1,928 人納入分析，平均年齡67 歲，女性居多 (65.5%)。接受TKR 手術占大部分 (68.4%)。術後多數病人並未使用抗凝血藥 (55.8%)，而有用藥者以warfarin 居多 (68.5%)。未用藥組病人的肝臟疾病 (1.8 vs. 0.4%, p = 0.007) 及腫瘤疾病 (5.9 vs. 3.5%, p = 0.023) 比例皆明顯高於用藥組。使用抗凝血藥組之VTE 發生率明顯高於未用藥組 (1.3 vs. 0.0%, p = 0.002)，使用warfarin 組VTE 發生率雖高於rivaroxaban 組 (1.5 vs. 0.7%, p = 0.339)，但未達統計差異性。未用抗凝血藥者其出血事件發生率明顯較用藥組高 (2.8 vs. 1.3%, p = 0.024)，使用warfarin 組出血事件發生率雖高於rivaroxaban 組，但亦無統計差異性 (1.4 vs. 1.1%, p = 0.759)。
結論：相較於其他文獻，本研究並未發現rivaroxaban 在療效與安全性方面有優於warfarin 之處。但本研究局限於單一醫學中心、醫師用藥習慣與加上病人出血風險高等因素，可能影響結果；因此rivaroxaban 用於關節術後栓塞的預防，值得進一步以大型資料庫探討
。

Objective: We investigated the efficacy and safety of rivaroxaban, warfarin, and no-medication groups for venous thromboembolism (VTE) prophylaxis after hip or knee arthroplasty.
Methods: A total of 1,928 patients with arthroplastic surgery of the knee (TKR) or hip (THR) during a 3-year period (2011 to 2014) with 6 months follow-up were enrolled in this retrospective cohort study. Those excluded were of ages < 18 years, with a history of thrombosis, atrial fibrillation, valve replacement, or with drugs switched between warfarin and rivaroxaban. The Kaplan-Meier analysis and log-rank test were used for statistical analyses.
Results: The mean age was 67.0 years old (65.5% female) with 68.4% received the TKR. More than half (55.7%) received no anticoagulant treatments (anticoagulants [ACs], or the ACs-free group), while the rest (44.2%, or the ACs group) received either warfarin or rivaroxaban. The ACs-free group, had higher rates of liver and cancerrelated diseases (liver: 1.8 vs. 0.4%, p = 0.007; cancer: 5.9 vs. 3.5%, p = 0.023). More VTE occurred in the ACs group (1.3 vs. 0.0%, p < 0.001). The warfarin subgroup had similar VTE incidences, comparing with the rivaroxaban subgroup (1.5 vs. 0.7%, p = 0.339). Bleeding incidences occurred more in the ACs-free group (2.8 vs. 1.3%, p = 0.024), and with slightly more bleedings in the warfarin subgroup (1.4 vs. 1.1%, p = 0.759).
Conclusion: This study was not able to demonstrate any significant difference in rivaroxaban and warfarin for thromboprophylaxis after arthroplasty. The results might due to limitations of this study. However, a large population study is needed to confirm the efficacy of rivaroxaban in Taiwan.