具有顱內出血病史之心房顫動病人的抗血栓藥物治療-回溯觀察性研究

Abstract

背景 雖然使在心房顫動的病人中會使用抗凝血劑來預防中風，但是在心房顫動且有顱內出血病史的病人中，是否該使用抗血栓藥物來預防中風，目前還沒有一個定論。因此本研究的目標是要探討在心房顫動且有顱內出血病史的病人中，現行台灣之治療方式以及使用抗血栓藥物的成效分析。 方法 本研究為一回朔觀察性研究，以「衛生福利部衛生福利資料科學中心」之全人口檔進行分析，並篩選出2011年至2017年中有心房顫動及顱內出血病史的病人。首先，會記錄發生顱內出血前後之抗血栓藥物使用情形和顱內出血後之使用抗血栓藥物時間點。再者，病人會依據顱內出血後第90天的藥物使用情形區分為：抗凝血藥物使用者、抗血小板藥物使用者以及無使用抗血栓藥物者。抗血栓藥物包含：warfarin, apixaban, dabigatran, edoxaban, rivaroxaban, aspirin, clopidogrel, ticagrelor和prasugrel。最後，病人會依據顱內出血後第一筆處方被區分為NOAC使用者以及Warfarin使用者。治療權重倒數機率方式校正抗血栓藥物使用者與非使用者的族群間差異，並利用Cox比例風險來評估兩個族群(OAC組vs No AT組, AP組 vs No AT組, NOAC組 vs warfarin組)中發生事件的風險。 結果 本研究共納入5007有心房顫動及顱內出血病史的病人。在發生顱內出血前，warfarin是最多被使用的(35.54%)，在發生顱內出血後，抗血小板藥物是最多被使用的(49.34%)，使用NOACs的比例從20.24%上升至25.75%，多於75%的病人在顱內出血後半年內開始使用抗血栓藥物。 第一部分研究中共納入283抗凝血藥物使用者、214抗血小板使用者以及1069無使用抗血栓藥物者。相較於無使用抗血栓藥物者，在治療意向分析法中，抗凝血藥物使用者有非顯著低的中風(HR: 0.583; 95%Cl:0.434-0.784)以及血栓事件風險(HR: 0.566;95%Cl:0.432-0.742)；並且在根據治療分析中，有顯著低的中風(HR: 0.612;95%Cl:0.422-0.888)以及血栓事件風險(HR:0.602;95%Cl:0.433-0.835)。抗凝血藥物使用者有顯著高的再發顱內出血(HR:2.260;95%Cl:1.399-3.653)以及無顯著差別的重大出血風險(HR:1.397:95%Cl: 0.987-1.979)；抗凝血藥物使用者治療意向分析法(HR: 0.943;95%Cl:0.828-1.074)和根據治療分析(HR: 0.851;95%Cl: 0.721-1.005)有非顯著低的全因死亡率。相較於無使用抗血栓藥物者，在治療意向分析法中，抗血小板藥物使用者有非顯著差別的中風(HR: 0.902;95%Cl:0.700-1.163)以及血栓事件風險(HR:0.993;95%Cl:0.792-1.245)；並且在根據治療分析中，有非顯著高的中風(HR:1.127;95%Cl:0.814-1.560)以及血栓事件風險(HR:1.113;95%Cl: 0.837-1.480)。抗血小板藥物使用者無顯著高的再發顱內出血(HR: 1.626;95%Cl: 0.985-2.683)以及重大出血風險(HR:1.102;95%Cl:0.769-1.580)；抗血小板藥物使用者治療意向分析法(HR:0.941;95%Cl:0.830-1.066)和根據治療分析(HR: 0.878; 95%Cl:0.747-1.033)有非顯著低的全因死亡率。 第二部分研究中共納入333 NOAC使用者以及205 warfarin使用者。相較於warfarin，在治療意向分析法中，NOACs有顯著低的中風(HR:0.605;95%Cl:0.377-0.971)以及血栓事件風險(HR:0.579;95%Cl:0.367-0.915)；並且在根據治療分析中，有非顯著低的中風(HR:0.923;95%Cl:0.502-1.697)以及血栓事件風險(HR:0.811; 95%Cl:0.459-1.435)。NOACs有顯著低的再發顱內出血(HR:0.332:95%Cl: 0.167-0.658)以及重大出血風險(HR:0.363:95%Cl:0.219-0.602)；NOACs在治療意向分析法(HR:0.662;95%Cl:0.518-0.846)和根據治療分析(HR:0.603;95%Cl:0.432-0.841)有非顯著低的全因死亡率。 結論 相較於無使用抗血栓藥物，抗凝血藥物比抗血小板藥物有較低的中風風險以及較高的出血風險，並且，抗凝血藥和抗血小板藥物在全因死亡率風險中均無顯著差別。在有心房顫動及顱內出血病史的病人中，NOACs是相較於Warfarin而言比較好的治療選擇。

Background Although anticoagulants are recommended in atrial fibrillation (AF) for stroke prevention, whether to start or resume antithrombotic therapy for AF patients surviving intracranial hemorrhage (ICH) remains debatable. The goal of our research was to document the treatment pattern and investigate the effectiveness and safety among antithrombotic (AT) therapy in AF patients with subsequent ICH in Taiwan. Methods A retrospective cohort study using the Full Population Data of the Health and Welfare Database was performed. AF patients with subsequent ICH from 2011 to 2017 were included. Firstly, the use of antithrombotic agents before and after the ICH, and time to initiation of antithrombotic agents after the ICH were documented. Secondly, Patients were divided into oral anticoagulant (OAC) user, antiplatelet (AP) agent users and AT agent non-users based on which they were prescribed with on the date after 90 days since the discharge date of the ICH. AT agents were defined as warfarin, apixaban, dabigatran, edoxaban, rivaroxaban, aspirin, clopidogrel, ticagrelor, and prasugrel. Thirdly, Patients were divided into NOACs users and warfarin users based on the first prescription after discharging from the ICH. IPTW was used to balance baseline characteristics. Cox proportional hazard ratios were used to evaluate the relationship between the therapeutic intervention (OAC users vs AT agent non-users, AP agent users vs AT agent non-users and NOACs users vs warfarin users) and outcomes of interests. Results 5007 AF patients with subsequent ICH were included in the study. Before ICH, warfarin therapy was the most used (35.54%). After ICH, antiplatelet therapy was most used (49.34%). The percentage of NOACs increased from 20.24% before the ICH event to 25.75% after the ICH event. More than 75% of the patients initiated AT agents within 6 months after discharging from the ICH. There were 283 OAC users, 214 AP agent users, and 1069 AT agent non-users. Compared to no AT therapy, OAC were significant lower in ischemic stroke (IS) (HR: 0.583;95%Cl:0.434-0.784), and thromboembolic event (TE) (HR:0.566;95%Cl:0.432-0.742) in intention-to-treat (ITT) analysis, and were significant lower in IS (HR:0.612; 95%Cl:0.422-0.888), and TE (HR:0.602;95%Cl:0.433-0.835). OAC were significant higher in recurrent ICH (HR:2.260;95%Cl:1.399-3.653. OAC were non-significant lower in all-cause mortality than AT agent non-users in ITT analysis (HR:0.943; 95%Cl:0.828-1.074) and AsT analysis (HR:0.851;95%Cl:0.721-1.005). AP agent were non-significant lower in IS (HR:0.902;95%Cl:0.700-1.163), and non-significant difference in TE (HR:0.993;95%Cl:0.792-1.245) in ITT analysis, and were non-significant higher in IS (HR:1.127;95%Cl:0.814-1.560), and TE (HR:1.113;95%Cl: 0.837-1.480) in AsT analysis. AP agent were non-significant higher in recurrent ICH (HR:1.626;95%Cl:0.985-2.683), and non-significant higher in MB (HR:1.102;95%Cl: 0.769-1.580). AP agent were non-significant lower in all-cause mortality in ITT analysis (HR:0.941;95%Cl:0.830-1.066) and AsT analysis (HR: 0.878; 95%Cl:0.747-1.033). There were 333 NOAC users and 205 warfarin users. Compared to warfarin, NOACs were significant lower in IS(HR:0.605;95%Cl:0.377-0.971) and TE (HR:0.579;95%Cl: 0.367-0.915) in ITT analysis, and were non-significant lower in IS (HR:0.923;95%Cl: 0.502-1.697), and TE (HR:0.811; 95%Cl:0.459-1.435) in AsT analysis. NOACs were significant lower in recurrent ICH (HR:0.332:95%Cl: 0.167-0.658) and major bleeding (HR:0.363:95%Cl:0.219-0.602). NOACs were significant lower in all-cause mortality in ITT analysis (HR:0.662;95%Cl:0.518-0.846) and in AsT analysis (HR:0.603; 95%Cl:0.432-0.841). Conclusions Compared to no AT agent, OAC had lower risk in stroke and higher risk in bleeding than AP agent, also, OAC and AP agent had no difference in all-cause mortality. NOACs should be preferred over warfarin for AF patients with subsequent ICH.