Efficacy and Safety of Andexanet Alfa for Bleeding Caused by Factor Xa Inhibitors: A Systematic Review and Meta-Analysis

Direct oral anticoagulants (DOAC) including factor Xa inhibitors are associated with bleeding events which can lead to severe morbidity and mortality. Reversal agents like andexanet alfa (AA) and 4F-PCC (Four-factor prothrombin concentrate complex) are available for treating bleeding that occurs with DOAC therapy but a comparison on their efficacy is lacking. In this study, we analyzed the efficacy and safety of patients treated with andexanet alfa for bleeding events from DOAC. Databases were searched for relevant studies where AA was used to determine efficacy and safety in bleeding patients who were on factor Xa inhibitors. Published papers were screened independently by two authors. RevMan 5.4 (The Cochrane Collaboration, 2020) was used for data synthesis. Odds ratio (OR) and mean difference (MD) was used to estimate the outcome with a 95% confidence interval (CI). Among 1245 studies were identified after a thorough database search and three studies were included for analysis. AA resulted in lower odds of mortality compared to 4F- PCC (OR, 0.37; 95% CI, 0.20-0.71) among patients with intracerebral hemorrhage. There was no difference in thrombotic events between patients receiving AA and 4F-PCC (OR, 2.40; 95% CI, 0.36-15.84). No differences in length of hospital stay and intensive care unit (ICU) stay were seen between patients receiving AA and 4F-PCC. In conclusion, andexanet alfa reduced in-hospital mortality in patients who had bleeding due to factor Xa inhibitors compared to 4F-PCC. However, there were no differences in thrombotic events, length of ICU, and hospital stay between patients treated with AA and 4F-PCC.


Introduction And Background
Direct oral anticoagulants (DOAC) have been increasingly used in patients for the prevention of systemic embolization in atrial fibrillation as well as treatment and prevention of deep vein thrombosis (DVT) and venous thromboembolism (VTE). As a result, the indications of DOAC have significantly expanded in the last decade [1][2][3][4][5]. Predictable pharmacokinetics and pharmacodynamics, rapid onset and offset of action, few drug interactions, and absence of need for regular laboratory monitoring provide an advantage to oral factor Xa inhibitors over traditional Vitamin K antagonists [6]. Factor Xa inhibitors also reduce fatal and intracranial hemorrhage compared with vitamin K antagonists [7,8]. However, fatal bleeding has been reported with oral factor Xa inhibitor use [8,9].
Before the introduction of andexanet alfa (AA), off-label use of 4 factor-prothrombin concentrate complex (4F-PCC) was advised and was used in the situation of life-threatening bleeding [10]. Prothrombin complex concentrates (PCCs) are isolated from fresh frozen plasma (FFP) and contain Vitamin K-dependent factors II, VII, IX, and X [11]. In May 2018, AA received FDA approval for use in patients treated with rivaroxaban and apixaban in the setting of life-threatening or uncontrolled bleeding following ANNEXA-A and ANNEXA-R trials in healthy participants [12,13]. AA is a modified recombinant, catalytically inactive form of human factor Xa, which binds and sequesters factor Xa inhibitor molecules that reduce anti-factor Xa activity rapidly in the body [14]. A multicenter, prospective, open-label, single-group study ANNEXA-4 was done in bleeding patients following FDA approval, which showed the drug's good efficacy and safety profile [15]. Randomized controlled trials have not been done, given the risks of using a placebo in acutely bleeding patients. However, some retrospective observational studies and case series studying the efficacy and safety of AA in bleeding patients have been published. In addition, some studies have compared efficacy and safety with 4F-PCC. We have conducted this systematic review and meta-analysis to analyze the effectiveness and safety profile of AA in bleeding caused by factor Xa inhibitors.

Review Methods
We used Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines for the systematic review of available literature [16]. The study protocol was registered in the International prospective register of systematic reviews (PROSPERO) CRD42021244219.

Literature search
We searched PubMed, PubMed Central, Scopus, Embase, and Cochrane library for relevant studies published till February 2021. Searches were conducted using the keywords like "andexanet alfa", "andexanet", "andexanet alpha", "bleeding", "factor Xa inhibitor," and "factor Xa inhibitors" and appropriate boolean operators. Details of the search strategy are available in Supplementary Material 1.

A. Types of Studies
We included studies done to determine the efficacy and safety of andexanet alfa in patients who had bleeding in the setting of factor Xa inhibitor use. As randomized controlled trials were not available, we included prospective and retrospective studies and case series with more than ten patients. AA was used to determine efficacy and safety in bleeding patients on factor Xa inhibitors in qualitative analysis. In addition, the studies with both treatment and control groups were included in the quantitative synthesis.

B. Types of Participants
The studies required patients to be more than 18 years of age and had bleeding in the setting of Factor Xa inhibitor use.

C. Types of Interventions
Andexanet alfa was taken in the treatment arm, while 4F-PCC or other blood products were included in the control arm.

D. Types of Outcome Measures
Our outcome of interest was hemostatic efficacy, mortality within 30 days, the incidence of thrombotic events, and length of hospital and ICU stay following treatment with AA or other blood products.
We excluded types of studies with the following characteristics: meta-analysis, reviews, in-vitro studies, studies done on healthy subjects, case reports, editorials, opinions, letters, protocols, abstracts/presentations, dissertation, and animal studies. Case series with fewer than ten patients, articles where full-text articles were not available, ongoing studies, and studies with incomplete data were also excluded.

Data extraction and management
Titles, abstracts, and full texts were screened for study and report characteristics that matched eligibility criteria. Studies were independently screened by two reviewers (AA and SS) using Covidence (Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia) and data were extracted for both quantitative and qualitative synthesis. The conflicts were resolved by taking the opinion of the third reviewer (NK). The data extraction sheet was created using Microsoft Excel software. One reviewer collected the data from all articles; the second reviewer verified the data for accuracy and highlighted discrepancies; the third reviewer resolved any disagreements and carried out a thorough evaluation to ensure that only the outcomes of interest were taken into account. The following variables were included: first author, type of design, site of study, year of publication, sample size, mean age, percentage of male and female, indication for anticoagulation, hemostatic efficacy, mortality within 30 days, length of hospital stay, length of ICU stay and incidence of thrombotic events.

Risk of Bias
We used the Joanna Briggs Institute (JBI) critical appraisal checklist for cohorts and case series for quality and risk of bias assessment ( Tables 1-2).

Statistical Analysis
RevMan 5.4 (The Cochrane Collaboration, 2020) was used for statistical analysis. Odds ratio (OR) and mean difference (MD) was used to estimate the outcome with a 95% confidence interval (CI).

Assessment of Heterogeneity
The statistical heterogeneity among the studies was calculated and assessed with the I 2 test based on previously recommended stratifications. In the case of heterogeneity, we used the invariance and random-effect finally, well. Finally, we evaluated the sensitivity by rerunning the analysis to assess any unrevealed differences.

Results
A total of 1245 studies were identified after thorough database searching, and 351 duplicates were removed. Title and abstracts of 894 studies were screened, and 860 irrelevant studies were excluded. The full-text eligibility of 34 studies was assessed, and 24 studies were excluded for definite reasons ( Figure 1). A total of 10 studies were included in the qualitative summary (Table 3), and three studies were included in the quantitative analysis.

Quantitative Analysis
Only three studies reported the use of AA contrasting with 4F-PCC among ICH patient groups used in synthesis.

Discussion
Our meta-analysis is the most comprehensive meta-analysis to evaluate the effect of andexanet alfa in bleeding caused by factor Xa inhibitors evaluating the mortality, length of hospital stay, length of ICU stay, and thrombosis in comparison to 4-F PCC. The major finding of our study was that andexanet alfa decreased mortality in patients who had intracerebral bleeding due to factor Xa inhibitors compared to 4F-PCC. There were 105 mortalities in 865 patients (12.13%) receiving andexanet alfa across ten studies. In contrast, the overall mortality rate in a recent 4F-PCC meta-analysis in FXa inhibitor bleeding was 18% compared to 12.13% in our analysis and 14% in the ANNEXA-4 trial [27]. The studies done by Ammar et al. and Barra et al. showed a higher mortality rate of 39% and 22% respectively which is higher than that of other studies as these studies included only ICH patients [17,18]. Mortality was also significant in a study done by Culbreth (40%) as 14 out of 15 patients had ICH. The Ammar et al. study showed a similar mortality rate in the andexanet group and 4F-PCC group (39% and 38% respectively) while the Barra et al. study showed higher mortality in the 4F-PCC group (63.6%) than in andexanet receiving patients (22.2%) [17,18]. Patients in the 4F-PCC group in the Barra et al. study had lower baseline GCS and higher baseline hematoma volume which might have contributed to the higher mortality [18].
We found no difference in the incidence of thrombotic events caused by AA in comparison to 4F-PCC for the reversal of bleeding caused by factor Xa inhibitor. A recent meta-analysis of seven studies including 240 patients showed thrombotic events of 4% with the use of 4F-PCC [27]. In contrast, we found 48 incidences of thrombosis among 523 patients in the nine studies included in our analysis. A prior meta-analysis done by Rodrigues et al. had estimated the risk of thrombosis with andexanet alfa and Idarucizimab at 5.5%, however, the analysis just included three studies for evaluation of thrombosis risk associated with AA and evaluated the cumulative risk of thrombosis associated with both andexanet alfa and Idarucizimab [28].  [22].
Clinical benefit of AA use was observed in bleeding due to factor Xa inhibitors in our analysis; however, the cost of stocking AA in most hospitals might be prohibitive for the immediate use for reversible DOAC related life-threatening bleeding. The median projected cost of andexanet alfa was $22,120/patient compared to $5670/patient for 4F-PCC. 4F-PCC currently is more widely available and less expensive, but that may change if the cost for AA comes down in the future [32]. 4F-PCC and andexanet alfa have not been compared in a prospective randomized clinical trial, and results of such studies are needed to inform clinical practice in DOAC related bleeding events. There is an ongoing randomized, multicenter clinical trial evaluating the efficacy and safety of andexanet alfa versus the usual standard of care in patients with ICH anticoagulated with a DOAC, which may be completed in 2023 [33].

Limitations of the study
Most of the studies included were case series and retrospective observational studies. Only one prospective study, the ANNEXA-4 trial, was included. There were control groups in only three of our studies which were all retrospective. The sample size was less in our studies. Therefore, there was a moderate to high risk of bias in our studies. ANNEXA-4 trial had wide exclusion criteria: planned surgery within 12 hours after andexanet alfa administration, ICH with GCS less than 7, hematoma volume more than 60 cc, expected survival less than one month, use of VKA, dabigatran, PCC, WB, or plasma in last seven days. Giovino's study also excluded patients with GCS less than 7 and hematoma volume >60 ml [24]. However, patients requiring surgical intervention, patients who received other blood products before AA administration, unknown time of the last factor Xa inhibitor dose, patients with low GCS and higher hematoma volume were included in other studies. In real clinical practice, patients with low GCS and expected mortality of less than one month required AA administration and were included in other studies. Knowledge about the administration of other blood products and time since the last factor Xa inhibitor was not feasible due to the retrospective nature of some studies and were thus included. Culbreth et al. 2019 included patients with bleeding due to factor Xa inhibitor who required emergent surgery.

Conclusions
Andexanet alfa reduced in-hospital mortality in patients who had bleeding due to factor Xa inhibitors compared to 4F-PCC. There was no difference in thrombotic events, length of ICU, and hospital stay between andexanet alfa and 4F-PCC. Thus, AA is a promising therapeutic agent for the reversal of factor Xaassociated bleeding. However, the cost of stocking AA in most hospitals might be prohibitive for the immediate use for reversible of DOAC related life-threatening bleeding. 4F-PCC currently is more widely available and less expensive, but that may change when the cost for AA decreases. More studies are required in the future to determine the effect of AA as compared to 4F-PCC in patients with DOAC-related bleeding other than intracranial bleeding.