Management of direct factor Xa inhibitor–related major bleeding with prothrombin complex concentrate: a meta-analysis

Siavash Piran, Rasha Khatib, Sam Schulman, Ammar Majeed, Anne Holbrook, Daniel M. Witt, Wojtek Wiercioch, Holger J. Schünemann and Robby Nieuwlaat


A targeted antidote for reversal of direct factor Xa (FXa) inhibitors is now available for clinical use in the United States, but it is costly and has limited availability. In a systematic review, we evaluated the safety and effectiveness of 4-factor prothrombin complex concentrate (4F-PCC) as an alternative for managing direct FXa inhibitor–related major bleeding. A systematic literature search was conducted using Medline, Embase, and the Cochrane Register of Controlled Trials up to September 2018. No comparative studies were found. Ten case series with 340 patients who received PCC for direct FXa inhibitor–related major bleeding were included. The pooled proportion of patients with effective management of major bleeding was 0.69 (95% confidence interval [CI], 0.61-0.76) in 2 studies using the International Society on Thrombosis and Haemostasis (ISTH) criteria and 0.77 (95% CI, 0.63-0.92) in 8 studies that did not use the ISTH criteria; all-cause mortality was 0.16 (95% CI, 0.07-0.26), and thromboembolism rate was 0.04 (95% CI, 0.01-0.08). On the basis of evidence with very low certainty from single-arm case series, it is difficult to determine whether 4F-PCC in addition to cessation of direct oral FXa inhibitor is more effective than cessation of direct oral FXa inhibitor alone in patients with direct FXa inhibitor–related major bleeding.


The options are limited for reversing the direct oral anticoagulant (DOAC) effect in patients who have major bleeding or who urgently need surgery. A targeted reversal agent for anti-factor Xa (anti-FXa) inhibitors, coagulation factor Xa (recombinant), inactivated is now available in the United States1,2 but it is costly and will initially be available only in a limited number of hospitals. Dabigatran is currently the only DOAC with the widely available direct antidote idarucizumab.3

Given the issues relating to coagulation factor Xa (recombinant), inactivated, nonspecific indirect reversal strategies are currently used in clinical practice, including prothrombin complex concentrates (PCCs), activated PCCs, and recombinant FVIIa.4 PCCs contain vitamin K–dependent coagulation factors and are widely used for reversal of vitamin K antagonists (VKAs).5 Four-factor PCCs (4F-PCCs) contain FII, FVII, FIX, and FX compared with 3-factor PCCs that lack FVII.6 PCCs have a lower risk of adverse events (volume overload in particular) than fresh frozen plasma (FFP), with no difference in risk of thromboembolic complications for warfarin reversal.7 Furthermore, PCCs have a smaller volume and can be administered more rapidly than FFP.5

The efficacy of PCCs in the reversal of direct FXa–related bleeding has been evaluated in healthy volunteers. 4F-PCCs corrected rivaroxaban-induced thrombin generation in 10 healthy volunteers in a dose-dependent manner.8 In 6 healthy nonbleeding volunteers who received rivaroxaban 20 mg twice per day for 2.5 days, a supratherapeutic plasma concentration was induced, but 4F-PCCs corrected the rivaroxaban-associated prolonged prothrombin time (PT) and increased endogenous thrombin generation potential.9 Two other studies in healthy volunteers receiving rivaroxaban 20 mg twice per day for 3 to 4 days to reach supratherapeutic steady-state concentrations suggested that 4F-PCCs normalized PT more effectively than 3-factor PCCs10 and that 4F-PCCs, but not tranexamic acid, reduced PT and increased endogenous thrombin potential.11 However, neither 4F-PCCs or tranexamic acid reduced bleeding duration or bleeding volume at the punch biopsy site.11 A phase 1 study evaluated the reversal of the effects of edoxaban on bleeding measures and biomarkers after punch biopsy by 4F-PCCs. 4F-PCCs given at doses of 50, 25, or 10 U/kg after administration of 60 mg edoxaban in a dose-dependent manner reversed the effects of edoxaban on bleeding duration and endogenous thrombin potential, with complete reversal seen at the 50 U/kg dose.12 Therefore, 4F-PCCs can effectively reverse thrombin generation; however, the effects on bleeding measures are variable as shown by the latter 2 studies in human volunteers of rivaroxaban at supratherapeutic doses and edoxaban at therapeutic doses.11,12

4F-PCCs have a proven efficacy in reversal of VKAs and variable results for reversal of direct FXa inhibitor coagulation assays in healthy human volunteers. In addition, because of high cost, lack of widespread availability of a specific reversal agent for direct FXa inhibitors, and a relatively safe adverse effect profile, 4F-PCCs may be a reasonable option for managing direct FXa-related major bleeding. It is not known whether the reversal effects of 4F-PCCs seen in healthy volunteers accurately reflect effects on clinical outcomes in patients with major bleeding, because no systematic review is available. We performed a systematic review and meta-analysis to address whether cessation of direct FXa inhibitor plus administration of 4F-PCCs vs cessation of direct FXa inhibitor alone can be used in patients with major bleeding while they are receiving apixaban, betrixaban, edoxaban, and rivaroxaban.


This systematic review was performed as part of the American Society of Hematology (ASH) clinical practice guidelines on venous thromboembolism (VTE), which were developed in partnership with the McMaster University Grading of Recommendations Assessment, Development and Evaluation (GRADE) Centre. Because the ASH guidelines focused on the specific topic of VTE, results and judgments for certainty of the evidence in this systematic review may be somewhat different than what is reported in the guidelines. Review and meta-analysis methodology followed the Cochrane Handbook,13 and reporting was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.14

Search strategy

We conducted a systematic literature search to identify potential studies in Medline (1996 to week 2 of September 2018), Embase (1974 to week 2 of September 2018), and the Cochrane Central Register of Controlled Trials (until week 2 of September 2018). The search strategy is provided in supplemental Table 1. We also reviewed the references of the included studies for additional potentially eligible studies. There were no restrictions based on language or publication status. The search was performed in blocks to examine the studies separately by design, including observational randomized controlled trials (RCTs) and systematic reviews. Narrative review articles, single case reports, letters to the editor, and editorials were excluded. We aimed to pool the results of these studies in a meta-analysis. The searches retrieved for each study type were combined at the end. Because we did not identify any systematic reviews or RCTs, we focused on observational studies.

Study selection

Two authors (S.P. and R.K.) independently screened the titles and abstracts of articles and then reviewed the full-length version of potentially relevant articlesfor eligibility. Disagreements were resolved by joint review and consensus. RCTs or nonrandomized studies (cohort, case-control, interrupted time series, case series) were eligible if they satisfied all of the following characteristics: involved adult patients (age ≥18 years) taking a direct FXa inhibitor (apixaban, betrixaban, edoxaban, rivaroxaban), evaluated the safety and effectiveness of 4F-PCCs in managing major bleeding, reported 1 or more outcomes such as thromboembolic events (including deep vein thrombosis and/or pulmonary embolism), stroke, myocardial infarction, major bleeding, and/or all-cause mortality.

Outcome measures

The effectiveness of major bleeding management was defined according to the International Society on Thrombosis and Haemostasis (ISTH) recommendation for effective management of major bleeding15 or defined as an absence of progression of hemorrhage on brain imaging by computed tomography in those presenting with intracranial hemorrhage, stabilization of bleeding on endoscopy, or cessation of surgical bleeding. Mortality was defined as death resulting from any cause. Thromboembolic events included deep vein thrombosis and/or pulmonary embolism, stroke, or myocardial infarction.

Risk of bias and indirectness assessment at the study level

Two authors (S.P. and R.K.) assessed the risk of bias using the Joanna Briggs Institute critical appraisal checklist for case series.16 We considered selection bias given that patients with severe bleeding were likely selected to receive a 4F-PCC and enrollment may not have been consecutive or random. We prespecified potential confounders such as age; alcohol intake; comorbidities such as a history of hypertension, chronic kidney disease, or liver disease; use of antiplatelet agents such as aspirin and/or clopidogrel; and potentially important cointerventions such as fresh frozen plasma, cryoprecipitate, platelets, FVIIa, and tranexamic acid. Although the presence of these confounding factors and cointerventions could not affect the comparison with other bleeding management approaches because only single-arm studies were included, they do provide information regarding how representative the included populations and their outcomes could be for our population of interest. We considered the information on confounding factors and cointerventions in the domain of indirectness.17

Certainty of body of evidence

We assessed the certainty of the body of evidence from all included studies in the domains of risk of bias, indirectness, imprecision, inconsistency, and publication bias using the GRADE approach.18 In accordance with the Cochrane Handbook,13 we could not use funnel plots to assess for publication bias given that we had <10 included studies in the systematic review. We then created a GRADE evidence profile using GRADEpro software (

Statistical analyses

Because the included studies were single arm only, we pooled the single-arm event rates using a random effects generic inverse variance method. We assessed heterogeneity using the I2 index, χ2 test, and visual inspection. Calculations were performed using RevMan version 5.3.


Study selection

The study selection process is shown in supplemental Figure 1. We identified 1410 citations in our literature search and found 58 potentially eligible studies. Forty-two studies were excluded: 33 were abstracts, 1 was a review article,19 4 articles were not related to DOAC-associated major bleeding,20-23 outcomes for patients treated with PCCs were not reported in 2 articles,24,25 and the bleeding outcome for each type of DOAC was unspecified in 8 studies.26-33 The remaining 10 studies were included in the systematic review and meta-analysis.

Patient and study design characteristics

The baseline characteristics of the included studies are listed in Table 1. A total of 340 patients were included in the 10 studies,34-43 all of which were single-arm case series (no comparison group) that included only patients who received 4F-PCCs for direct FXa inhibitor–related major bleeding. Follow-up duration had a median of 9 days in 1 study,40 30 days in 3 studies,38,39,42 90 days in 5 studies,34,35,37,41,43 and 180 days in 1 study.36 In 9 studies, atrial fibrillation was the most common indication for a direct oral FXa inhibitor therapy. One study did not specify the indication for anticoagulation.41 Most patients were receiving rivaroxaban at the time of bleeding except for 107 patients who received apixaban. Cointerventions were used in 9 studies, including FFP, platelets, or recombinant FVIIa. Two of the 10 studies excluded patients who had received FFP.35,42 None of the patients received activated PCCs.

Table 1.

Baseline characteristics of the studies included in the systematic review

Risk of bias and certainty of evidence

All the included studies were rated as having very serious risk of bias because they lacked a control group and had a high risk of confounding and selection bias. Only 2 studies included consecutive patients (supplemental Table 2).38,39 Furthermore, outcome assessors in all 10 studies were not blinded. The risk of evidence imprecision was rated as very serious, given that the included studies had small sample sizes and very few events. Overall, the certainty of evidence was rated as very low (Table 2).

Table 2.

Assessment of quality of evidence using GRADE approach


Outcomes of interest are listed in Table 3. Two studies used the ISTH effective major bleeding management criteria,38,39 and 8 studies did not.34-37,40-43 Therefore, a pooled meta-analysis of the effectiveness of management of all 10 studies was not performed. All of the studies except for 141 reported mortality outcomes and were included in the pooled crude proportion analysis for overall mortality. Three studies did not report thromboembolic events,36,37,41 and we did not include these in the pooled crude proportion analysis for thromboembolic complications.

Table 3.

Outcomes of the studies included in the systematic review

Pooled analysis of outcomes

Forest plots of the pooled outcome proportions are provided in supplemental Figures 2-5. The pooled proportion of effective major bleeding management using the ISTH criteria was 0.69 with a 95% confidence interval (CI) of 0.61 to 0.76, meaning that 69% of patients achieved successful bleeding management using 4F-PCCs. This was pooled from 2 studies involving 150 patients.38,39 The pooled proportion of effective management of major bleeding that did not use the ISTH criteria was 0.77 (95% CI, 0.63-0.92). This was pooled from 8 studies involving 190 patients.34-37,40-43 The crude mortality pooled proportion was 0.16 (95% CI, 0.07-0.26), meaning that 16% of patients died during the specified follow-up period. This was pooled from 9 studies involving 249 patients.34-40,42,43 The crude pooled proportion of thromboembolic events was 0.04 (95% CI, 0.01-0.08), meaning that 4% of patients experienced a thromboembolic event during follow-up. This was pooled from 7 studies involving 240 patients.34,35,38-40,42,43 In a subanalysis, we separated the effect on short-term thromboembolic complication only for 5 studies that had a follow-up of ≤30 days (supplemental Figure 6).38-40,42,43 The short-term pooled proportion of thromboembolic complications was 0.03 (95% CI, 0.0-0.06). There was no statistical heterogeneity between the studies for most outcomes (I2 = 0%) except for effective management of major bleeding not using the ISTH criteria (I2 = 85%) and crude mortality (I2 = 74%).


We found 10 nonrandomized, unblinded, single-arm case series in which all patients received 4F-PCC for management of direct FXa-related major bleeding. Therefore, these data were pooled to determine the proportion of patients who experienced benefits and harms after receiving 4F-PCC. Because there was no comparator group, it was difficult to determine whether administration of 4F-PCC in addition to cessation of direct oral FXa inhibitor was more likely to achieve hemostasis and improve important outcomes than cessation of direct oral FXa inhibitor alone.

Our study is the first systematic review to examine the safety and effectiveness of 4F-PCC for managing major bleeding in patients receiving FXa inhibitors to treat VTE or atrial fibrillation. The search was systematic and comprehensive and consisted of searches in multiple electronic databases. Our study has limitations. First, the included studies had a high risk of bias and small sample sizes. All were case series that lacked a control group and likely suffered from selection bias. Only 2 studies minimized the selection bias by including consecutive patients.38,39 In addition, there may also be a risk of confounding biases because of differences in the baseline characteristics or cointerventions between study populations. For example, cointerventions such as tranexamic acid, FFP, platelets, or recombinant FVIIa were used in all studies, which could have influenced the bleeding outcome. The certainty of evidence was rated very low because of the high risk of bias and imprecision resulting from small sample sizes and few events. Furthermore, we could not assess publication bias using funnel plots because we had <10 studies. In addition, there was a high degree of heterogeneity for the pooled outcomes of effective major bleeding management not using the ISTH criteria and crude mortality. This is most likely because of differences in baseline characteristics in the individual studies. For example, in 1 study, a tertiary care center received patient referrals from other hospitals in which previous supportive measures had failed.39 Second, we combined patients with different types and doses of 4F-PCC. Finally, the definition of outcome events was not uniform between studies. The recent ISTH criteria for effective management of major bleeding were used in only 2 studies.15

Management of major bleeding according to the ISTH criteria was effective in 103 (69%) of the 150 patients in 2 studies. This is similar to the results of the ANNEXA-4 trial: coagulation factor Xa (recombinant), inactivated reduced the anti-FXa activity in patients with acute major bleeding and achieved good or excellent hemostasis in 109 (83%) of 132 patients.2 Without a control group, it is unclear whether coagulation factor Xa (recombinant), inactivated is more effective in achieving hemostasis than supportive care alone. Because of the high cost and limited availability of coagulation factor Xa (recombinant), inactivated, many clinicians may continue to choose 4F-PCC for managing direct FXa-related major bleeding.

The crude pooled proportion of thromboembolic complications in patients who received 4F-PCCs in our meta-analysis (4%) was higher than the incidence of 1.8% for 4F-PCC-associated thromboembolic complication for managing VKA-related bleeding32 but far below that of the reported rate when using coagulation factor Xa (recombinant), inactivated (11%).2 The difference may be explained by differences in baseline characteristics, the timing of re-initiation of anticoagulation after administering the reversal agent, the intrinsic properties of the reversal agent, or the combination. Transient elevations in D-dimer and prothrombin fragments 1 + 2 levels were observed in healthy volunteers after administration of andexanet alfa,44 which is thought to be mediated by interaction between andexanet and tissue factor pathway inhibitor. As with the case series included in our meta-analysis, a controlled study is required to determine whether the reported rate of thromboembolic events exceeds the expected rate in patients who are already at increased risk of thromboembolic complications at baseline.

The crude all-cause mortality pooled proportion was higher than that reported in the ANNEXA-4 trial (18% vs 12%, respectively).2 The mortality rate observed in our meta-analysis is likely influenced by selection bias whereby sicker populations were selected across studies. Furthermore, 251 (74%) of the 340 patients included in our meta-analysis had intracranial hemorrhage (ICH) compared with 61% of the patients in the ANNEXA-4 study.2 Anticoagulant-related ICH is generally known to have a high mortality rate.45-48 The mortality rate in patients with warfarin-related ICH treated with PCC has been reported as 45%,49,50 which is higher than the rate observed in our study (25.5% [36 of 141 ICH patients in 8 studies]). A recent observational study of 42 patients with oral anticoagulant–associated ICH (28 taking VKAs and 14 taking an FXa inhibitor) were treated with 4F-PCC.42 Although the sample size is small, the outcomes were similar in both groups, with no significant difference in in-hospital mortality (32.1% in the VKA group vs 14.2% in the FXa inhibitor group) or rates of hemorrhagic expansion, thromboembolic complications, or length of stay.42

In conclusion, our systematic review and meta-analysis is the first of its kind to examine safety and effectiveness outcomes of 4F-PCC for managing direct FXa inhibitor–related major bleeding. However, in light of the absence of a comparator group, it is impossible to know whether 4F-PCC is more effective than supportive care alone. For example, if there is no change in hematoma volume in patients with ICH, without a control group it cannot be determined whether this would have occurred without administering 4F-PCC. In addition, gastric bleeding from deep ulcers with a visible blood vessel may not be amenable to treatment with 4F-PCC or coagulation factor Xa (recombinant), inactivated. For such bleeding types, invasive interventions are required. The rate of thromboembolic events must be taken into consideration when evaluating the harms and benefits associated with administration of 4F-PCC. Given the quick offset of DOAC anticoagulant effects in patients with adequate renal function and a low but clinically relevant rate of thromboembolic complications after administration of 4F-PCC, clinicians should reserve this intervention for life-threatening bleeding only. Because of the high cost and lack of widespread availability of a specific reversal agent for direct FXa inhibitors, 4F-PCC may be a reasonable option for managing direct FXa-related major bleeding. Because of the high risk of bias and small sample sizes across studies, the quality of the evidence is very low. Future studies with a control group are needed to determine whether reversal with 4F-PCC is more effective than supportive care alone.


This systematic review was performed as part of the ASH clinical practice guidelines on VTE. The entire guideline development process was funded by ASH.


Contribution: S.P. contributed to study design, study selection, collection of data, statistical analyses, and writing the initial draft of the manuscript; R.K. contributed to study design, study selection, and collection of data; S.S. and A.M. interpreted the data and provided vital reviews of the manuscript; and A.H., D.M.W., W.W., H.J.S., and R.N. contributed to study design, interpreted the data, and provided vital reviews of the manuscript.

Conflict-of-interest disclosure: A.M. and S.S. are lead authors of 2 of the studies included in the systematic review. The authors declare no competing financial interests.

Correspondence: Siavash Piran, Thrombosis Service, HHS General Hospital, 237 Barton St East, Hamilton, ON L8L 2X2, Canada; e-mail: siavash.piran{at}


  • The full-text version of this article contains a data supplement.

  • Submitted July 27, 2018.
  • Accepted November 3, 2018.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
View Abstract