Ecacy and Safety of Bivalirudin Versus Heparin Anticoagulation Therapy for Extracorporeal Membrane Oxygenation: Meta-Analysis

We aimed to compare the ecacy and safety of bivalirudin versus heparin as the anticoagulant in patients undergoing Extracorporeal Membrane Oxygenation (ECMO). We conducted a search in PubMed, Embase and the Cochrane Library for all the studies in which bivalirudin was compared to heparin as the anticoagulant for ECMO. Ecacy outcomes were dened as the time to reach therapeutic levels, time within therapeutic range (TTR), thrombotic events, circuit thrombosis, circuit exchanges. Safety outcomes were reported as Heparin-Induced Thrombocytopenia (HIT), major bleeding events, minor bleeding events. Other outcomes included hospital length of stay (LOS), ICU LOS, mortality, 30-day mortality and in-hospital mortality. minor bleeding events[OR 0.93, 95%CI 0.38,2.29, p=0.87, I 2 =0%], hospital LOS[MD -2.93, 95%CI -9.01,3.15, p=0.34, I 2 =45%], ICU LOS[MD -4.22, 95%CI -10.07,1.62, p=0.16, I 2 =0%], mortality[OR 1.84, 95%CI 0.58,5.85, p=0.30, I 2 =60%] and 30-day mortality[OR 0.75, 95%CI 0.38,1.48, p=0.41, I 2 =0%]. The benet of bivalirudin over heparin was not signicant for patients undergoing ECMO for major bleeding events while ruling out the Rivosecchi’s study (OR 0.44, 95%CI 0.71-1.14). Subgroup analysis by patient’s type revealed that studies in children generated lower rate of thrombotic events and major bleeding events compared with adults.


Introduction
Extracorporeal membrane oxygenation (ECMO) is widely used for the circulatory and respiratory support [1] . Anticoagulant is an essential component for patients undergoing extracorporeal membrane oxygenation [2] . Thrombosis events and bleeding events are common complications [3] . Unfractionated heparin (UFH) remained to be the primary anticoagulant for ECMO in guidelines due to ease of titration and monitoring, ease of reversibility, low cost [4] . Despite these advantages, heparin has its limitations.
First, heparin require the cofactor antithrombin III for e cacy [5] . Second, it may cause heparin-induced thrombocytopenia due to platelet dysfunction and its highly antigenic, with mortality as high as 20%-30% [6] . Third, it only inhibits free thrombin.
In recent years, bivalirudin, a direct thrombin inhibitor (DTI), has been used as an alternative for patients requiring ECMO [7] . As a DTI, Bivalirudin shows the following advantages. First, bivalirudin does not require the antithrombin III for e cacy as it binds directly to thrombin, allowing for more consistent effect. Second, it does not cause the occurrence of HIT. Third, it inhibits both circulating and clot-bound thrombin. Although some data regarding to the use of bivalirudin as an anticoagulant in ECMO have been published, the reports have been limited. This meta-analysis will review bivalirudin anticoagulation strategies in ECMO patients.

Search
We did an electronic search from January 1, 2010 to September 1, 2021 of the following databases: PubMed, Embase, and the Cochrane Library. The keywords "Bivalirudin", "Heparin" and "Extracorporeal membrane oxygenation" were searched. Two investigators (JG and HJ Y) independently screened the titles and abstracts to ascertain whether each study met the eligibility criteria. The full texts of the identi ed eligible articles were then evaluated to determine whether they should be included in the analysis. Disagreements between the two reviewers were resolved by consensus.

Selection criteria
Inclusion criteria were a) the study was prospectively or retrospectively designed; b) patients were included in the heparin or bivalirudin if they received solely heparin or bivalirudin; We excluded studies that a) reported outcomes with only heparin or only bivalirudin; b) lacked data detailing the outcomes included in our analysis.

Data abstraction and quality appraisal
Surname of the rst author, year of publication, country of origin, study period, study design, type, number of patients, study group, targeted ACT/APTT, age, gender, ECMO type and ECMO duration were extracted for each potentially included study. The data extraction was conducted by two independent investigators (JG and HJ Y). Any discrepancy was solved by discussion and intervention of a senior investigator. The validity of included studies was appraised with the Newcastle-Ottawa scale [8] .

Data analysis
Treatment effects were expressed as odds ratios for binary outcomes and mean difference for quantitative outcomes. Betweenstudy heterogeneity was assessed using I 2 statistic and P value. The xed-effect model was applied if no or low signi cant heterogeneity was present. To explore heterogeneity, we did subgroup analyses and sensitivity analyses. All statistical analyses were conducted with RevMan software (version 5.3) and Stata software (version 14.0). A two-sided P value<0.05 was considered statistically signi cant.

Demographic characteristics and Quality assessment
The characteristic of studies reporting bivalirudin versus heparin as an anticoagulant for ECMO patients are summarized in Table 1 and Table 2. 10 studies reporting on 1091 patients treating with bivalirudin or heparin as an anticoagulant for ECMO patients. Bivalirudin was administered in 405 patients while 686 patients were treated with heparin. All studies were published after 2011. All studies were Retrospective non-randomized clinical trial. There were 2 studies from Italy, 8 studies from American. The mean/median age and the sex of the patients were extractable in 8 studies with the mean/median age ranging from 12 months to 56.8 years; 589(61.4%) of the patients were male. 551 patients received veno-arterial (VA) ECMO and 407 veno-venous (VV) ECMO. The mean/median ECMO duration ranged from 106 hours to 10 days (Table 1/Table 2). The quality assessment is displayed in eTable 1.

Primary meta-analysis
The Bivalirudin regimens during ECMO The loading dose of bivalirudin was not administered in our systematic review. The maintenance infusion dosages of bivalirudin range from range from 0.01mg/kg/h to 0.5mg/kg/h, even with similar targeted APTT or ACT. APTT and ACT were reported in 8 and 1 studies respectively. The targeted APTT ranged from 45-90s, while the targeted ACT ranged from 160-180s (Table 1).

Sensitivity analysis and subgroup analyses
The bene t of bivalirudin over heparin was not signi cant for patients undergoing ECMO for major bleeding events while ruling out the Rivosecchi's study (OR 0.44, 95%CI 0.71-1.14) (eFigure 1). Subgroup analysis by patient's type revealed that studies in children generated lower rate of thrombotic events and major bleeding events compared with adults (eFigure 2).

Discussion
To our knowledge, our study was the rst meta-analysis review which address the e cacy of safety of bivalirudin compared with heparin in patients undergoing ECMO. The major ndings from our systematic review of 10 observational studies are as follows: First, bivalirudin bolus or infusion dose varied signi cantly between studies, which range from 0.01mg/kg/h to 0.5mg/kg/h. Second, compared with heparin, bivalirudin demonstrated e cacy and safety for systematic anticoagulation on extracorporeal membrane oxygenation as established by a decrease in the number of thrombotic complications (circuit thrombosis), bleeding events and hospital mortality. Third, there was no signi cant differences between groups regarding the time to reach therapeutic levels, TTR, circuit exchanges, HIT, minor bleeding, hospital LOS, ICU LOS, mortality and 30-day mortality.
Challenges we encountered while developing the optimal antithrombotic regimen during ECMO. To our knowledge, there is no consensus on bolus dosing or infusion dosing of bivalirudin. Therapeutic APTT was achieved without the bolus dose in our meta-analysis. The maintenance infusion rates varied signi cantly between studies, which range from 0.01mg/kg/h to 0.5mg/kg/h, even with similar targeted APTT or ACT. Previous review found that the maintenance dose ranged from 0.045 to 0.48mg/kg/h in children and 0.025 to 0.05 mg/kg/h in adults, which is similar to our results [19] . However, Other studies demonstrated poor correlation with coagulation tests including APTT [20] . Patients' variability and variability of the APTT between laboratories based on reagents may play a role in the dosing strategy. Further studies are required to assess the correct bivalirudin bolus or infusion dose for ECMO treatments.
The time to reach therapeutic levels and TTR represent directly to the quality of the anticoagulation dose management. Our systematic review does not indicate that bivalirudin exhibited a more consistent APTT control over time compared with heparin, which may be primarily due to small sample size. Although bivalirudin does not have signi cant more consistent APTT value compared with heparin in our study, it appears to have an impact in patient outcomes including thrombotic complications and major bleeding events. However, it should be noticed that the bene t of bivalirudin over heparin was not signi cant for major bleeding events in sensitivity analysis. First, the de nitions for bleeding events varied, which contributed to high heterogeneity among the trials. Second, the meta-analysis might be dominated by a single large study, which led to the insigni cant results of major bleeding events while removing the large study. Third, Other cause attributed to this result may be the varied dose of initiation and titration increment dose in different studies.
Heparin depends on the cofactor antithrombin III for e cacy while bivalirudin binds directly to thrombin. Heparin resistance is more frequent in children than in adults due to lower concentration of antithrombin in neonates or critically ill children [21] .
Children have de ciencies in anticoagulant hemostasis proteins due to liver immaturity [22] . For these reasons, the disadvantages of using heparin are more pronounced in children. On our subgroup analyses based on the patient's type, bivalirudin remained to be associated with deceased thrombotic events and major bleeding events compared with heparin. Meanwhile, the superiority of bivalirudin over heparin was more evident in children compared with adults.
Our meta-analysis has several limitations. First, potential bias is likely to be greater for observational studies. Therefore, the results should be interpreted with caution due to methodological heterodetic. Second, the de nitions of adverse outcomes differed between studies, which should account for the variance between studies. The standardization of de nitions for outcomes will allow for a more consistent reporting of outcomes in future clinical investigations.

Conclusions
Overall, Bivalirudin can be used safely and effectively to decease thrombotic complications, bleeding events and in-hospital mortality when compared with heparin for patients undergoing ECMO. Bivalirudin may offer a clinically signi cant advantages as the anticoagulant of choice. Our ndings support the hypothesis, but the pooled results should be interpreted with caution that the results of sensitivity analysis were not consistent. The superiority of bivalirudin over heparin for anticoagulation in the ECMO population requires further prospective randomized controlled studies.

Declarations
Availability of data and materials All data generated and/or analyzed during the current study are included within the published article and its additional ies.
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable. na. not available; NRCT non-randomized controlled trial  Figure 1 Flow chart of the search process Forest plot for the time to reach therapeutic levels, TTR, thrombotic events, circuit thrombosis and circuit exchanges Figure 3 Forest plot for HIT, major bleeding events and minor bleeding events