Effects of different anticoagulant drugs on the prevention of complications in patients after arthroplasty: A network meta-analysis

Abstract

Background: After arthroplasty treatment, some complications commonly occur, such as early revision, infection/dislocation, and venous thromboembolism (VTE). This study aims to use a network meta-analysis to compare effects of 9 anticoagulant drugs (edoxaban, dabigatan, apixaban, rivaroxaban, warfarin, heparin, bemiparin, ximelagatran, and enoxaparin) in preventing postoperative complications in arthroplasty patients.

Methods: After retrieving PubMed, Embase, and Cochrane Library database from the inception to November 2016, randomized controlled trials were enrolled. The integration of direct and indirect evidences was performed to calculate odd ratios and the surface under the cumulative ranking curves. Nineteen eligible randomized controlled trials were included.

Results: The network meta-analysis results showed that compared with warfarin, edoxaban, apixaban, and rivaroxaban had a lower incidence rate in asymptomatic deep venous thrombosis, which indicated that edoxaban, apixaban, and rivaroxaban had better effects on prevention. Similarly, in comparison to enoxaparin, edoxaban and rivaroxaban had better effect; rivaroxaban was better than ximelagatran in preventive effects. Compared with apixaban, edoxaban, dabigatan, rivaroxaban, and enoxaparin had a higher incidence rate in clinically relevant non-major bleeding, which showed that preventive effects were relatively poor. In addition, the results of the surface under the cumulative ranking curves showed that rivaroxaban and bemiparin worked best on symptomatic deep venous thrombosis and pulmonary embolism. In terms of bleeding, apixaban and warfarin had better preventive effects.

Conclusion: Our findings suggested that rivaroxaban may work better in terms of symptomatic deep venous thrombosis and pulmonary embolism, whereas apixaban had better preventive effects in bleeding.

Figure 1

Cochrane system bias evaluation of all the nineteen enrolled studies in this network meta-analysis.

Figure 2

Funnel plot for publication bias of included studies. (A) Edoxaban; (B) dabigatan; (C) apixaban; (D) rivaroxaban; (E) warfarin; (F) heparin; (G) bemiparin; (H) ximelagatran; (I) enoxaparin. ADVT = asymptomatic deep venous thrombosis, CRNM = clinically relevant non-major, PE = pulmonary embolism, SDVT = symptomatic deep venous thrombosis.

Figure 3

The evidence network of all enrolled studies about the preventive effect of the 9 anticoagulant drugs (edoxaban, dabigatan, apixaban, rivaroxaban, warfarin, heparin, bemiparin, ximelagatran, and enoxaparin) on the ADVT and CRNM bleeding in postoperative patients receiving arthroplasty in this network meta-analysis. ADVT = asymptomatic deep venous thrombosis, CRNM = clinically relevant non-major.

Figure 4

Forest plots for the relationship of the effect of different anticoagulant drugs on the prevention of postoperative ADVT in patients after arthroplasty. (A) Edoxaban; (B) dabigatan; (C) apixaban; (D) rivaroxaban; (E) warfarin; (F) bemiparin; (G) ximelagatran; (H) enoxaparin. ADVT = asymptomatic deep venous thrombosis.

Figure 5

Forest plots for the relationship of the effect of different anticoagulant drugs on the prevention of postoperative CRNM bleeding in patients after arthroplasty. (A) Edoxaban; (B) dabigatan; (C) apixaban; (D) rivaroxaban; (E) enoxaparin. CRNM = clinically relevant non-major.