Background The Fontan circulation is associated with an increased risk of thromboembolic events (TEs). As many as 25% of these thrombotic events result in fatality. More subtle adverse effects on the pulmonary circulation from embolic thrombi may further impair adequate functioning of the circuit. Despite these well-documented phenomena, the most optimal approaches to thromboprophylaxis are still not clearly defined.
Objective A meta-analysis of published trials in English on PubMed and Cochrane libraries that evaluated the role of using TE prophylaxis in patients who underwent the Fontan procedure was conducted.
Methods 10 studies with a total number of 1200 patients with an average follow-up time of 7.1 years were identified. A random effect model was used.
Results The incidence of TE was significantly less in patients who received TE prophylaxis (using either aspirin or warfarin) compared with patients who did not receive TE prophylaxis (OR 0.425, 95% CI 0.194 to 0.929, p<0.01, I2=37%). The incidence of TE was significantly lower in patients who received aspirin compared with no TE prophylaxis (OR 0.363, 95% CI 0.177 to 0.744, p<0.01, I2=0%) and who received warfarin compared with no TE prophylaxis (OR 0.327, 95% CI 0.168 to 0.634, p<0.01, I2=2.5%). There was no significant difference in incidence of TE between warfarin and aspirin (OR 0.936, 95% CI 0.609 to 1.438, p=0.54, I2=0%). Furthermore, there was no significant difference in incidence of early TE (within 6 months of the operation) or late TE (>6 months) between patients receiving warfarin and aspirin (OR 0.784, 95% CI 0.310 to 1.982, p=0.37, I2=8%) and (OR 0.776, 95% CI 0.249 to 2.42, p=0.3, I2=45%), respectively. When only total cavopulmonary connection patients were included, there was again no difference between warfarin and aspirin in the incidence of TE (OR 0.813, 95% CI 0.471 to 1.401, p=0.34, I2=11%).
Conclusions This study shows a significantly lower incidence of TE after Fontan procedure if either aspirin or warfarin is used. This meta-analysis suggests no significant difference in incidence of early or late TE in patients receiving aspirin compared with warfarin.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Since it was first reported in 1968, the Fontan procedure has transformed the lives of many children born with single-ventricle physiology, offering them the potential for survival and good quality of life well into adulthood.1 ,2 The original procedure was described by Fontan in patients with tricuspid atresia, and since that time, the surgical technique has significantly evolved.2
To overcome the many haemodynamic challenges of the original Fontan design, the total cavopulmonary connection was developed, which has a significantly more streamlined design and avoids right atrial dilation as in the earlier versions of the circulation.3
Despite all these modifications, thrombosis (both arterial and venous) has remained a major complication following the Fontan procedure and carries a reported fatality of up to 25%.1 ,4 The presentation may be intracardiac or extracardiac including pulmonary embolism, cerebrovascular thromboembolism, deep venous thrombosis or other embolic phenomena.1 ,5 Potential risk factors for thrombosis include not only haemodynamic factors with low cardiac output and less pulsatile pulmonary flow, but also well-documented abnormalities in coagulation.6 ,7 Several studies have shown that patients with Fontan circulation carry decreased circulating levels of key anticoagulants, such as protein C and protein S.6 Whether this is secondary to liver dysfunction or to ongoing consumption of these proteins remains unclear. Other studies have shown that patients with Fontan circulation have evidence of prothrombotic endothelial activation, including increased circulating levels of von Willebrand's factor, factor VIII and soluble thrombomodulin, as well as platelet activation.8 ,9
Strategies to prevent thrombosis and thromboembolism (TE) after the Fontan procedure vary considerably among institutions, and the literature is fraught with controversy.1 ,2 ,10 While few studies discourage the use of any TE prophylaxis, some studies support the use of antiplatelet agents and others recommend the use of therapeutic anticoagulation with vitamin K antagonists.2 ,7 ,11 ,12
The incidence of TE seems to be higher in the first six months after Fontan surgery, and, hence, many institutions use vitamin K antagonists during the first six months following surgery, followed by indefinite antiplatelet therapy.13 Concerns regarding the potential role of the progressive thrombophilia that develops in most patients with Fontan circulation over time in the late failure of the Fontan circulation raise important questions as to whether warfarin should be administered in preference to antiplatelet therapy alone. The relative long-term benefit of warfarin over aspirin, if any, particularly after the new generation Fontan procedure, remains unclear at present. Therefore, a systematic review and meta-analysis of the available literature was performed to establish the most optimal method of TE prophylaxis after the Fontan procedure.
A meta-analysis of published studies that evaluated the role of using TE prophylaxis in patients who underwent the Fontan procedure was conducted. Both randomised and retrospective studies were identified. Only studies that made a direct comparison between different strategies of TE prophylaxis or TE prophylaxis versus no prophylaxis were included. PubMed and the Cochrane Library for trials published in English were searched (figure 1). Two independent investigators did the study search and agreed on the studies included. The search strategies are included in the online supplementary file. The risk of bias was assessed using Cochrane risk of bias tool for randomised controlled trial and modified Newcastle Ottawa scale for retrospective studies (see online supplementary Funnel plot and tables).
The meta-analysis was conducted using a random-effect model, but the results of the fixed-effect model analysis were also reported for comparison. MedCalc (V.15.6) software was used to conduct the analysis. A computerised model was used. Under this model, weights were assigned to all studies based entirely on the amount of information captured by that study (sample size and event occurrence). The primary outcome was the diagnosis of TE that was evident clinically or by different imaging modalities.
As the incidence of TE may differ between the new generation and the old generation Fontan, a subgroup analysis for patients who underwent the new generation Fontan procedure (total cavopulmonary connection) was performed. Only two studies reported the incidence of TE in old generation Fontan. In one of these studies, patients were taking warfarin only versus no treatment while in the second study patients were taking warfarin or aspirin versus no treatment, and thus, the subgroup analysis was not performed.11 ,14 Finally, the effect of the timing of the surgery on the choice of TE prophylaxis was studied. TE events that occurred within 6 months of surgery were considered early events, while those that occurred after 6 months were considered late events.
Ten studies that met the inclusion criteria were identified (figure 1). These encompassed a total of 1200 patients with an average follow-up time of 7.1 years15–18 (table 1). Among the 1200 Fontan patients reported, the overall TE incidence was 11.3%. Patients who received TE prophylaxis with either aspirin or warfarin had lower incidence of TE in patients post-Fontan compared with no prophylaxis (OR 0.425, 95% CI 0.194 to 0.929, p<0.01, I2=37%) (figure 2). The overall incidence of TE in patients receiving no prophylaxis was 18.6% compared with 8.6% when aspirin was used and 9% when warfarin was used (table 2). There was also a statistically significant lower TE incidence in patients receiving aspirin compared with no TE prophylaxis (OR 0.363, 95% CI 0.177 to 0.744, p<0.01, I2=0%), as well as warfarin compared with no TE prophylaxis (OR 0.327, 95% CI 0.168 to 0.634, p<0.01, I2=2.5%) (figure 2). There was, however, no significant difference between warfarin and aspirin in TE incidence (OR 0.936, 95% CI 0.609 to 1.438, p=0.54, I2=0%) when all types of Fontan procedure were considered, regardless of the timing of the event after surgery.
When only new generation Fontan was considered in the analysis (seven studies were included), there was still no significant difference in TE incidence between patients treated with warfarin versus aspirin (OR 0.813, 95% CI 0.471 to 1.401, p=0.34, I2=11%) (figure 3). There was not enough data on the old generation Fontan operation to enable a direct comparison of TE prophylaxis strategies within this group as only two studies reported these results.
There was also no significant difference between patients receiving warfarin versus aspirin in TE incidence early (within 6 months of the operation) (OR 0.784, 95% CI 0.310 to 1.982, p=0.37, I2=8%) or late (>6 months) (OR 0.776, 95% CI 0.249 to 2.42, p=0.3, I2=45%) (figure 3).
Of note, when aspirin dose was reported it was administered at 3–5 mg/kg/day. Aspirin resistance testing of any kind was not used in any of the studies to guide treatment. The international normalised ratio (INR) goal for patients on warfarin therapy was 2–3 in all studies with an exception of Cheung et al,14 who reported a goal of 1.5–2.5. One study reported an increase in the risk of thrombosis in the warfarin group in patients who had subtherapeutic INRs compared with patients with therapeutic INRs.5
This meta-analysis shows an overall TE incidence of 11.3% after the Fontan procedure. There was a lower incidence of TE in patients receiving prophylaxis (either aspirin or warfarin) compared with no TE prophylaxis. There was no significant difference between warfarin and aspirin in incidence of TE. Patients who received aspirin or warfarin had lower TE incidence by >50% compared with no TE prophylaxis. However, failure rates were still relatively high (∼9%) with either agent. When only the new generation Fontan was considered, there was still no difference between warfarin and aspirin in TE incidence and failure rates remained significantly high with either agent. Furthermore, there was no significant difference between warfarin and aspirin in TE incidence either early (within 6 months of the operation) or late (>6 months) after the Fontan procedure.
Despite the use of pharmacological prophylaxis and the new generation Fontan procedure, TE incidence remains high in this patient population. Moreover, TE accounts for significant morbidity and mortality following Fontan surgery.1 Indeed, previous studies have demonstrated that mortality significantly increases in patients who underwent Fontan surgery after a single TE.1 ,19 Khairy et al1 showed that TE is the cause of death in 25% of the Fontan patients who died during a follow-up period of 12.1 years post-Fontan procedure.
The incidence of TE post-Fontan varies greatly between studies. This variability may, at least in part, be due to the differences in the method of TE diagnosis between studies. Some studies relied solely on the presence of symptomatic events or routine transthoracic echocardiograms to identify TE, while other studies used more invasive methods, including trans-oesophageal echocardiogram and chest CT angiogram.5 Trans-oesophageal echocardiography has a significantly greater detection rate for intracardiac thrombosis in Fontan patients, and this has been consistently documented in the literature.20
The aetiology of the thrombophilic state associated with the Fontan procedure is likely multifactorial. Several studies have shown a relative imbalance between the procoagulant and anticoagulant proteins leading to a hypercoagulable state after the Fontan procedure. The levels of protein C, protein S and antithrombin were found to be low while the level of factor VIII increases significantly after the Fontan procedure.6 ,7 Whether the diminution in levels of protein C, protein S and antithrombin is due to the local consumption of these key anticoagulant proteins at local sites of subclinical thrombosis and/or poor production due to liver dysfunction remains to be determined. Factor VIII is an acute-phase reactant and may be elevated in these patients simply as a result of systemic stress due to organ dysfunction. Alternatively, elevated factor VIII levels may represent a sign of endothelial dysfunction, a conclusion suggested by several studies.8 ,9 Other studies have also shown evidence of increased platelet reactivity in patients with Fontan circulation, suggesting that the thrombophilia observed in these patients is not simply a function of abnormal coagulation.21 ,22 In fact, the findings of primary platelet hyper-reactivity in this context led to studies comparing aspirin and warfarin to prevent TE after the Fontan procedure.
The importance of TE prophylaxis has been emphasised in many studies included in this meta-analysis.10 ,23 However, there was no consensus between the studies in terms of the preferred prophylaxis strategy. Most of the studies, as shown in figure 2, showed no difference between aspirin and warfarin with respect to the incidence of TE. The exception is the study by Manlhiot et al,23 which showed a decreased incidence of TE in patients who were discharged home on warfarin after the Fontan procedure. However, the same group published a multicentre randomised-controlled trial that showed no difference between the two prophylactic strategies in patients who underwent extracardiac conduit Fontan procedure.5
New generation Fontan procedures were thought to be associated with a lower risk for TE in older reports.3 Other studies showed no difference in the incidence of TE with different types of the Fontan procedure.24 When analysis for patients with new generation Fontan only was done, there was no difference between aspirin and warfarin in terms of the number of TE events. There was insufficient data to compare warfarin to aspirin in the old generation Fontan.11 ,14 ,25
Warfarin has many limitations, including narrow therapeutic range, delayed onset and offset of action, numerous drug and food interactions, significant pharmacogenetic variations and the need for an intensive management approach to ensure effective and safe anticoagulation. Aspirin is also not without limitations. Aspirin resistance has been reported to occur in 10% of patients with congenital heart disease and up to 52% of adults with Fontan circulation.26 While there is no clear benefit to warfarin over aspirin for preventing TE in patients with Fontan circulation, one cannot make the general conclusion that targeting platelet function is as efficacious as anticoagulation in preventing Fontan-associated TE.7 Time spent in the therapeutic range of warfarin is highly variable in clinical practice and was not frequently reported in the clinical trials except in one study that showed higher incidence of TE in patients with lower INR.5 The highly variable pharmacokinetics of warfarin and the fact that it also significantly diminishes levels of key vitamin K-dependent anticoagulants (ie, proteins C and S) prompt the need to study alternative agents that overcome these barriers.27 Whether the newer oral anticoagulants with superior pharmacokinetics and very high specificity that have been shown to be excellent alternatives to warfarin in other contexts would be used in preventing Fontan-associated TE remains to be determined.28 ,29 Adult data suggest that the benefit of these new agents compared with warfarin is marginal in patients with greater time in the therapeutic ranges.30 Data from McCrindle et al31 also suggest that the patients who spend more time in therapeutic range have less risk of having TE in Fontan patients. Despite these observations, the clinical conundrum remains, however, that many patients on warfarin, even when in a trial setting, have significant periods outside of the therapeutic range, and perhaps this is not surprising given the heterogeneity of this patient population and the anticipated associated pharmacodynamic variation.
There is evidence that the risk of TE is highest within the first few months after surgery.23 ,31 Due to this increased risk, the practice of many institutions is to use warfarin for the first six months after the Fontan procedure, then to transition to aspirin.7 ,13 To further investigate the effectiveness of this practice, TE events were divided into early (within first six months postoperatively) and late (beyond six months postoperatively). No significant difference between aspirin and warfarin in incidence of early or late TE events was found. When comparing TE prophylaxis strategies between both groups, the possibility of bleeding should be taken into consideration. Bleeding complications were not reported in most of the studies reviewed. However, the literature regarding warfarin use generally demonstrates a relatively low bleeding risk of <1–2% per year.23 ,32–35
The coexistence of different risk factors increases the incidence of thrombosis after the Fontan procedure. These risk factors include anatomic and haemodynamic factors in addition to the prolonged use of central lines following the Fontan procedure, atriopulmonary type of Fontan connection, a history of previous thrombosis and the material of the extracardiac conduit.7 ,12 ,19 ,24 ,31 Risk factors were not reported in most of the studies reviewed and were not included in the analysis.
All but one of the trials included are observational non-randomised trials, with only one multicentre randomised-controlled trial. Thus, this meta-analysis inherits the limitations of non-randomised trials. Additionally, since most studies were retrospective and some of them chose to give warfarin based on thrombosis risk factors, selection bias is a possible limitation to our analysis.36 Also, for the old generation Fontan circulation, there was insufficient data to perform the analysis. Despite these limitations, we believe that this meta-analysis does provide valid insights into effectiveness of aspirin versus warfarin in managing thrombophilic risk in this heterogeneous group of patients.
In this meta-analysis, TE incidence was found to be lower when prophylaxis (with aspirin or warfarin) was used post-Fontan procedure. These data suggest no significant difference between warfarin and aspirin in incidence of TE early and late post-Fontan procedure. The relatively high incidence of TE in patients with Fontan circulation despite treatment with either aspirin or warfarin speaks to significant confounding factors in their use. These factors include aspirin resistance and the logistic challenges of maintaining INR in the therapeutic range for majority of time. Clearly, research strategies to overcome these limitations are necessary and exploration of other treatment modalities in the management of these complex thrombophilic states associated with the Fontan circulation is indicated.
What is already known on this subject?
Fontan circulation is associated with an increased risk of thromboembolic events (TEs), which cause a significant morbidity and mortality. Strategies to prevent TE after the Fontan procedure vary considerably among institutions.
What might this study add?
This meta-analysis shows an overall TE incidence of 11.3% after the Fontan procedure. Patients who received aspirin or warfarin had lower TE incidence by >50% compared with no TE prophylaxis. There was no significant difference between warfarin and aspirin in incidence of TE.
How might this impact on clinical practice?
This study confirmed the importance of TE prophylaxis in the Fontan population. This study showed that aspirin use for primary TE prophylaxis seems to be as effective as warfarin and has many advantages over warfarin in terms of pharmacodynamics and drug and dietary interactions.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Data supplement 1 - Online supplement
- Data supplement 2 - Online supplement
- Data supplement 3 - Online supplement
Contributors All authors take responsibility for the content of the manuscript and have approved the manuscript and agree with submission to Heart. TA and SA wrote the first draft of the manuscript. TA, SA and JF reviewed the included studies in the meta-analysis. CCA critically revised the manuscript. JSP wrote the haematology part of the manuscript, and GRV is the senior author.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.