Efficacy of Prostacyclin Anticoagulation in Critically Ill Patients Requiring Extracorporeal Support: A Systematic Review and Meta-Analysis

Extracorporeal support modalities are highly prothrombotic. Anticoagulation is frequently used for patients receiving Continuous Renal Replacement Therapy (CRRT), Molecular Adsorbent Recirculating System (MARS), and Extracorporeal Membrane Oxygenation (ECMO). The objective of this systematic review and meta-analysis is to determine if prostacyclin-based anticoagulation strategies are effective compared to other anticoagulation strategies, in critically ill children and adults who needs extracorporeal support, such as continuous renal replacement therapy. We conducted a systematic review and meta-analysis using multiple electronic databases and included studies from inception to June 1, 2022. Circuit lifespan, proportion of bleeding, thrombotic, and hypotensive events, and mortality were evaluated. Out of 2,078 studies that were screened, 17 studies (1,333 patients) were included. The mean circuit lifespan was 29.7 hours in the patients in the prostacyclin-based anticoagulation series and 27.3 hours in the patients in the heparin- or citrate-based anticoagulation series, with a mean difference of 2.5 hours (95%CI -12.0;16.9, p=0.74, I2=0.99, n=4,003 circuits). Bleeding occurred in 9.5% of the patients in the prostacyclin-based anticoagulation series and in 17.1% of the patients in the control series, which was a statistically significant decrease (LogOR -1.14 (95%CI -1.91;-0.37), p<0.001, I2=0.19, n=470). Thrombotic events occurred in 3.6% of the patients in the prostacyclin-based anticoagulation series and in 2.2% of the patients in the control series, which was not statistically different (LogOR 0.97 (95%CI -1.09;3.04), p=0.35, I2=0.0, n=115). Hypotensive events occurred in 13.4% of the patients in the prostacyclin-based anticoagulation series and in 11.0% of the patients in the control series, which was not statistically different (LogOR -0.56 (95%CI -1.87;0.74), p=0.40, I2=0.35, n=299). The mortality rate was 26.3% in the prostacyclin-based anticoagulation series, and 32.7% in the control series, which was not statistically different (LogOR -0.40 (95%CI -0.87;0.08), p=0.10, I2=0.00, n=390). The overall risk of bias was low to moderate. In this systematic review and meta-analysis of 17 studies, prostacyclin-based anticoagulation was associated with fewer bleeding events, but with similar circuit lifespans, thrombotic events, hypotensive events, and mortality rates. The potential benefits of prostacyclin-based anticoagulation should be explored in large randomized controlled trials.


Introduction And Background
Thrombotic complications are a frequent concern in patients receiving Continuous Renal Replacement Therapy (CRRT) [1], Molecular Adsorbent Recirculating System (MARS) [2], and Extracorporeal Membrane Oxygenation (ECMO) [3]. CRRT is a life support modality used for acute kidney injury, while ECMO is used for severe respiratory and cardiac failure. Both therapies involve the use of extracorporeal circuits in which blood comes in contact with artificial surfaces, which increases the risk of thrombosis. To prevent these complications, anticoagulation is typically administered [1,4], which should ideally minimize clot formation while also minimizing the risk of bleeding [5].
Most CRRT anticoagulation strategies involve systemic unfractionated heparin or regional citrate infusions, while most ECMO anticoagulation strategies involve unfractionated heparin or bivalirudin [1,5]. These strategies target the coagulation cascade but fail to address another part of hemostasis, the platelet function.
Prostacyclin or prostaglandin I2 (PgI2), a prostaglandin member of the eicosanoid family of polyunsaturated fatty acids, inhibits platelet activation and is an effective vasodilator. While prostacyclin is widely used to treat pulmonary hypertension, its antiplatelet effect is seldom sought [6]. Since this molecule has a different mechanism than heparin, citrate, or bivalirudin, its benefit-to-risk ratio might be valuable in critically ill patients requiring extracorporeal support. However, there is limited evidence regarding its safety, efficacy, and impact on patient outcomes [7,8] in this population.
The objective of this systematic review and meta-analysis is to compare the efficacy of prostacyclin-based anticoagulation strategies (either PgI2 alone or in association with another anticoagulant) with other anticoagulation strategies in critically ill patients requiring extracorporeal support.

Design
This is a systematic review of observational studies or randomized trials to determine the duration of the extracorporeal circuit, the prevalence of bleeding, thrombosis, thrombocytopenia, hypotension, and mortality, according to the anticoagulation strategy (prostacyclin as compared to other anticoagulation strategies or no intervention), in critically ill children and adults who needs extracorporeal support. The review followed recommendations contained within the PRISMA statement [9]. This study was registered with PROSPERO (Record ID: 332326).

Types of Studies
We included all studies, including abstracts submitted to conferences, case series, retrospective or prospective controlled studies, and randomized controlled trials that 1) enrolled critically ill children and/or adults who require extracorporeal support, such as continuous renal replacement therapy (CRRT), molecular adsorbent recirculating system (MARS), and extracorporeal membrane oxygenation (ECMO); 2) that compared the use of prostacyclin with other anticoagulation strategies; and 3) that report at least one outcome of interest. We included "grey literature", which includes abstracts of unpublished studies and studies published in non-peer-reviewed journals. We excluded studies where prostacyclin was used as a pulmonary vasodilator, studies without a comparator (i.e. prostacyclin only), studies without clinical outcomes (e.g. laboratory values only), studies outside of a critical care setting (bypass, chronic dialysis, etc.), individual case reports, case series of 10 patients or less, reviews or editorials not providing original results, and animal studies.

Types of Participants
We included studies that enrolled subjects on extracorporeal support (CRRT, MARS, and/or ECMO). Patients were further categorized as children or adults.

Types of Interventions
Based on the use of prostacyclin, anticoagulation strategies were categorized as intervention (either PgI2 alone or in association with another anticoagulant) or as control, when the anticoagulation strategy did not include PgI2 (e.g. heparin, citrate, or no anticoagulation).

Types of Outcome Measures
Our primary outcome was circuit lifespan (in hours). Our secondary outcomes were the proportion of bleeding events, thrombotic events, thrombocytopenia, hypotensive events, and mortality rate, as defined by the authors of each study.

Search Methods for Identification of Studies
For this systematic review, we performed a search in Ovid MEDLINE, Ovid Embase, and Cochrane from the inception of the database to June 1, 2022. The search included keywords and controlled vocabulary for prostacyclin, anticoagulation, and extracorporeal support (CRRT, MARS, ECMO) (Supplemental Tables 3-5). We imported the results to Covidence (Covidence Systematic Review Software, Veritas Health Innovation, Melbourne, Australia), which automatically detected duplicates.

Selection of Studies
Five reviewers (NA, AAA, MA, MJ, and OK) independently examined all potential studies and decided on their inclusion in the review (Figure 1). We independently evaluated each study based on its methods and outcomes. We performed this process without blinding of the study authors, institutions, journals of publication, or results. We resolved disagreements through additional evaluation by another author.

Data Extraction and Management
For each study included in the systematic review, two authors independently extracted data. We resolved disagreements through discussion. If required, we contacted study authors to ask for relevant data (e.g. specifying the definition of bleeding or the standard deviation of the average duration of circuit life). The corresponding authors were emailed twice, within a two-week period.

Assessment of Risk of Bias in Included Studies
We evaluated the validity and design characteristics of each study looking for aspects of major potential biases (study participation, prognostic factor measurement, outcome measurement, study confounding, and statistical analysis) [10]. Two authors reviewed and ranked each study's quality factor separately and defined studies as having a low risk of bias only if they adequately fulfilled all the criteria.

Assessment of Effect Size
We reported the duration of the extracorporeal filter as mean and standard deviation (SD), and proportions and their 95% confidence interval (95% CI) as the number of patients with the outcomes of interest (bleeding, thrombosis, thrombocytopenia, hypotension, or death) over the total number of patients enrolled in the study. The effect size for continuous variables was the mean difference, whereas the effect size for binary outcomes was the log of the odds ratio. To account for small proportions and interval confidences close to 0, we pooled the individual proportions using arcsin transformation [11].
Assuming each study estimated a study-specific true effect, we used random-effect models to pool odds ratios. Such models assume no a priori knowledge about the association between the real, or apparent, intervention effects; the differences between the studies are considered to be random. These models account for heterogeneity, with the center of this distribution describing the average of the effects, and its width describing the degree of heterogeneity. We used the DerSimonian-Laird random-effect method in the presence of significant heterogeneity [12].
We explored the effects of the age categories (adults vs children), the risk of bias (low risk, moderate risk, and high risk), and experimental anticoagulation strategy (prostacyclin alone vs prostacyclin and another anticoagulant), using Harbord's random-effect meta-regression model [13].

Assessment of Heterogeneity
We explored heterogeneity using the I2 statistic. An I2 statistic higher than 50% represents substantial heterogeneity [14].

Subgroup Analysis
We provided subgroup analyses of outcomes that were significantly different between both arms.

Funnel Plot
We explored funnel plot asymmetry as a surrogate for potential publication biases, solely in randomized controlled trials with significant findings, as the performance of funnel plots in observational studies has not been thoroughly validated [15].

Statistical Analyses
Mean differences (MD), proportion with 95% CI, and random-effects model were undertaken using SPSS (IBM SPSS Statistics for Windows, Version 28.0. Armonk, NY: IBM Corp.).

Description of Studies
We identified a total of 2,148 references, of which 70 were duplicates and therefore removed from review, leaving a total of 2,078 studies that were screened. Of these, 2,031 studies were irrelevant, leaving 47 fulltext articles to be further assessed. Of the 47 article, 25 were excluded because of a non-critical care setting, absence of original data, fewer than 10 patients receiving prostacyclin, duplicates between an abstract and a full text, etc. (Figure 1). In addition, we reached out to the corresponding authors of seven studies, two of which provided the requested data. Five did not respond within the predefined time or declined to provide additional information and were excluded since none of the outcomes of interest were reported in a standardized way (e.g. only the number of circuit changes but not the duration of circuit lifespan was reported). Finally, 17 studies met the eligibility criteria [16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32]. Of these, seven were only available as abstracts [16,21,[24][25][26]29,30].
Thirteen studies were observational while four were randomized controlled trials [17,22,23,28]. Sixteen studies enrolled 1261 patients on continuous renal replacement therapy (CRRT) and one study enrolled 72 patients on MARS [25]. No studies enrolled patients on ECMO.
The studies are described in Table 1
A random-effect meta-regression model to evaluate the effect of co-variables showed that circuit lifespan was not associated with quality of evidence (p=0.83) and prostacyclin strategy (p=0.81). As all these studies enrolled adults, the effect of the age could not be evaluated in this regression model.
A random-effect meta-regression model evaluating the effect of co-variables showed that the proportion of bleeding events was not independently associated with quality of evidence (p=0.49), age category (p=0.55), or prostacyclin strategy (p=0.95). Figure 4 is the subgroup forest plot comparing the two prostacyclin strategies (i.e. prostacyclin alone vs with heparin).
The funnel plot performed for the three randomized control trials that reported bleeding outcome [17,23,28] is symmetrical, which doesn't suggest publication bias ( Figure 5).
Using a random-effect meta-regression model to evaluate the effect of co-variables, we observed that the proportion of hypotensive events was not independently associated with age categories (p=0.41) or prostacyclin strategy (p=0.91). As all but one study had a low risk of bias, the effect of the quality of evidence could not be evaluated in this regression model.
Using a random-effect meta-regression model to evaluate the effect of co-variables, we observed that the mortality rate was not independently associated with age categories (p=0.62) or prostacyclin strategy (p=0.45). As all but one study had a low risk of bias, the effect of the quality of evidence could not be evaluated in this regression model.

Discussion
This meta-analysis of 17 studies, that enrolled 1,333 patients requiring extracorporeal support (mostly CRRT), showed a significant reduction in bleeding events in the prostacyclin-based strategy when compared to other strategies. There were no statistically significant differences in circuit lifespan, hypotensive events, thrombotic events, thrombocytopenia, and mortality.
Circuit failure, which is defined as a premature failure of the circuit before the expected lifespan, might occur because of the complex blood-circuit biocompatibility, patient's critical condition, inflammation, size and position of vascular access, and efficiency of anticoagulation [33]. Circuit failure might lead to suboptimal treatment efficacy [34], and increased transfusion requirements [35]. Although the available literature about the economic impact of premature circuit failure is scarce, a shorter circuit lifespan is likely associated with higher costs. While our results failed to show an improved circuit lifespan, they do suggest a decreased risk of bleeding. Bleeding complications can be secondary to the patient's underlying coagulopathy, complications of vascular access, or secondary to the need for anticoagulation. While heparin-induced bleeding has been reported to be as high as 30% on CRRT [36], our meta-analysis showed the risk of bleeding to be 17% in the control groups (heparin or citrate), and only 9.5% in patients who received prostacyclin. Therefore, prostacyclin-based anticoagulation strategies could decrease bleeding events without increasing thrombotic events or decreasing the circuit lifespan.

Limitations
While this meta-analysis is the first to evaluate prostacyclin in all types of extracorporeal support and in both critically ill adults and children, certain limitations must be acknowledged. First, most of the included studies were observational in nature, which might have led to a selection bias. Second, as for all systematic reviews, there might have been a publication bias. Third, only two studies enrolled pediatric patients [26,30]. Fourth, there was heterogeneity among the intervention, since some studies used a prostacyclin dose of 1 ng/kg/min [23] to 10 ng/kg/min [27]. In addition, the co-anticoagulation was also heterogeneous, with some studies using heparin at various doses, and others using citrate. Fifth, the outcome definitions were also heterogeneous, as some studies defined circuit failure based on different cutoff trans-membrane pressures, while others used visible circuit clots. Finally, most of the studies enrolled patients on CRRT in which blood flow rate and circuit characteristics factor in CRRT failure and circuit clotting (apart from MARS that uses a primary CRRT device).This prevents us from extrapolating these results to other modalities, such as ECMO or plasma exchange therapy.

Conclusions
In conclusion, in a systematic review and meta-analysis of 17 studies, prostacyclin-based anticoagulation was associated with fewer bleeding events, but with similar circuit lifespans, thrombotic events, hypotensive events, and mortality rates. Our study has some limitations, such as heterogeneity in the interventions (prostacyclin alone vs prostacyclin with other anticoagulants) and in the outcome definitions. In addition, nearly all studies enrolled patients on continuous renal replacement therapy, which prevents our results to be extrapolated to other modalities of extracorporeal support, such as extracorporeal membrane oxygenation. The potential benefits of prostacyclin-based anticoagulation should be further explored in large randomized controlled trials to investigate the optimal dose as well as the applicability to other extracorporeal support modalities.

Appendices
Search Strategy

Conflicts of interest:
In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.