The Fontan-circulation is the palliative approach for patients born with a congenital cardiac defect in which only one ventricle is capable to sustain the systemic circulation. Systemic venous return is redirected to the pulmonary circulation in a 2-step approach by connecting the superior caval vein to the pulmonary artery-system (see module Glenn-shunt). The final step to create the Fontan circulation is to connect the inferior caval vein to the pulmonary circulation, either by a lateral tunnel, partly incorporating atrial tissue in the Fontan circulation, or by placing a rigid Gortex tube between the ICV and the pulmonary arteries. In both approaches a fenestration can be made between the Fontan-pathway and the atrium, creating a right-to-left shunt.

The risk of thrombosis is increased after Fontan completion with reported incidences of venous thrombosis ranging from 4 to 19% and the incidence of stroke ranging from 3 to 19% in children with thrombo-prophylaxis (Attard, 2018). Of note, the risk of thrombosis seems to be highest during the first year after Fontan completion (Attard, 2018).

Several pathophysiological mechanisms play a role in the increased risk of thrombosis in single ventricle patients with a Fontan circulation, including endothelial dysfunction, abnormal blood flow, hypercoagulability (Attard, 2018).

Because of the increased risk for thrombosis and thromboembolic events there is consensus that single ventricle patients with a Fontan circulation need some form of thromboprophylactic therapy, in which decreasing the risk of thromboembolic events outweighs the risk of increased bleeding. However, in single ventricle patients with a Fontan circulation several thromboprophylactic regimens were studied and used but it is currently unclear which thromboprophylactic strategy is most favorable for children (Van den Eynde, 2023).

1. Studies with follow-up ≤ 1 year

Table 3. Summary of Findings – Comparison of thromboprophylaxis A versus thromboprophylaxis B with outcomes thrombosis, hemorrhage, mortality, ischemic stroke, quality of life (QOL) and adverse events

Population: Children with an indication for the Fontan procedure

Intervention: Thromboprophylaxis A (specified in table)

Comparator: Thromboprophylaxis B (specified in table)

Click here to see both tables in a document

_{^1. Risk of bias: some concerns (-1 level). No exclusion criteria were defined, and it is unclear whether the outcomes of interest were not present at the start of the study. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^2. Risk of bias: some concerns (-1 level). No exclusion criteria were defined, and it is unclear whether the outcomes of interest were not present at the start of the study. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^3. Risk of bias: some concerns (-1 level). No exclusion criteria were defined, and it is unclear whether the outcomes of interest were not present at the start of the study. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^4. Risk of bias: some concerns (-1 level). Method of randomization is unclear, no reporting on allocation concealment, and the study was not blinded. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^5. Risk of bias: some concerns (-1 level). Method of randomization is unclear, no reporting on allocation concealment, and the study was not blinded. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^6. Risk of bias: some concerns (-1 level). Method of randomization is unclear, no reporting on allocation concealment, and the study was not blinded. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^7. Risk of bias: some concerns (-1 level). Method of randomization is unclear, no reporting on allocation concealment, and the study was not blinded. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^8. Risk of bias: some concerns (-1 level). Method of randomization is unclear, no reporting on allocation concealment, and the study was not blinded. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^9. Risk of bias: some concerns (-1 level). Method of randomization is unclear, no reporting on allocation concealment, and the study was not blinded.}

2. Studies with follow-up > 1 year

Table 4. Summary of Findings – Comparison of thromboprophylaxis A versus thromboprophylaxis B with outcomes thrombosis, hemorrhage, mortality, ischemic stroke, quality of life (QOL) and adverse events

Population: Children with an indication for the Fontan procedure

Intervention: Thromboprophylaxis A (specified in table)

Comparator: Thromboprophylaxis B (specified in table)

_{^1. Risk of bias: some concerns (-1 level). It is unclear whether the outcomes of interest were not present at the start of the study, confounder adjustment is not consistently conducted across all study, method of randomization is unclear, no blinding, and no reporting on allocation concealment. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^2. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, unclear reporting on multivariable statistical analyses, method of randomization unclear, and no blinding. Imprecision: serious (-1 level). The confidence interval crosses one border of clinical relevance.}

_{^3. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study and unclear reporting on multivariable statistical analyses. Imprecision: serious (-1 level). The confidence interval crosses one border of clinical relevance.}

_{^4. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, unclear reporting on multivariable statistical analyses, method of randomization is unclear, and no blinding. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^5. Risk of bias: some concerns (-1 level). Method of randomization is clear, no blinding, and no reporting on allocation concealment. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{6. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^7. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^8. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^9. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^10. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^11. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^12. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups.}

_{^13. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^14. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^15. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^16. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^17. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^18. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^19. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups.}

_{^20. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^21. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

_{^22. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups.}

_{^23. Risk of bias: some concerns (-1 level). It is unclear whether the outcome of interest was not present at the start of the study, no proper confounder adjustment was conducted, co-interventions differed between the intervention groups. Imprecision: very serious (-2 levels). The confidence interval crosses both borders of clinical relevance.}

Description of studies

A total of ten studies were included in the analysis of the literature. Important study characteristics and results are summarized in table 2. The assessment of the risk of bias is summarized in the risk of bias tables (under the tab ‘Evidence tabellen’).

1. Studies with follow-up ≤ 1 year

Ankola (2021) conducted a single-center retrospective cohort study in patients who underwent the Fontan procedure between January 1, 2006 and December 31, 2016. No further inclusion and exclusion criteria were defined.

McCrindle (2021) conducted an RCT in children with single-ventricle physiology who had the Fontan procedure within four months before enrollment. Further inclusion criteria were age between 2 and 8 years of age and being clinically stable and able to tolerate or nasogastric feedings. Exclusion criteria were having thrombosis, a history of gastrointestinal disease or surgery associated with impaired absorption, active bleeding or high risk of bleeding contraindicating antiplatelet or anticoagulant therapy, including history of intracranial hemorrhage, or contraindications to aspirin or rivaroxaban.

2. Studies with follow-up > 1 year

Egbe (2017) conducted a retrospective cohort study in adult patients (> 18 years) with a history of a Fontan operation at Mayi Clinic, Rochester, Minnesota, from January 1994 to June 2014 who were identified from the hospital’s patient database using free-text search software. Patients with fewer than 12 months’ follow-up after the beginning of the study period were excluded.

Kawamatsu (2021) conducted a retrospective cohort study at three centers specializing in the care of adult patients with congenital heart disease in Japan: St. Luke’s International Hospital, Chiba Prefectural Cardiovascular Center, and Tsukuba University Hospital. Patients over 15 years of age with Fontan circulation who visited any of the three institutions from April 2015 to March 2018 were included. No exclusion criteria were defined.

Al-Jazairi (2019) conducted a retrospective cohort study in patients who underwent Fontan procedures and were identified from the Congenital Heart Defects Registry of the King Faisal Specialist Hospital and Research Centre. Included patients underwent Fontan procedures that took place at King Faisal Specialist Hospital and Research Centre, Riyadh, between 1985 and 2010. Exclusion criteria were patients who died, were lost to follow-up, or had a Fontan redo within one year of the first Fontan procedure.

Iyengar (2016) conducted a cohort study in children who underwent the Fontan procedure, (specifically the extracardiac conduit procedure) and were registered in the Australia and New Zealand Fontan registry. Exclusion criteria were death prior to discharge, patients referred from outside Australia or New Zealand, having mechanical prosthetic valves, patients who were transplanted or died during the first year after the Fontan procedure, patients who were lost to follow-up for any period of more than five years, and patients with < 1 year of follow-up available.

Pessotti (2014) conducted a single-center RCT in children who underwent total cavopulmonary operation with extracardiac conduit in the period between 2008 and 2011. Inclusion criteria were (1) children with heart with univentricular morphology in staged programming, for total cavopulmonary shunt operation. Exclusion criteria were (1) some aspect of anatomical or angiographic features that contraindicated the cavopulmonary operation, including a) main ventricular ejection fraction less than 60%, or significant impairment of atrioventricular valve related to the predominant ventricle, (b) serious anatomical alteration of the pulmonary venous return (untreatable), (c) anatomy of severely unfavorable pulmonary tree, (2) Inability to perform outpatient treatment, (3) any medical condition that prevented the randomization, for example, any indication that the patient used necessarily one of the two drugs evaluated, or contraindication to the use of them, (4) refusal of legal guardian to sign the Post-information Consent.

Seipelt (2002) conducted a cohort study in patients who underwent modified Fontan procedure for various types of univentricular heart conditions. No further inclusion and exclusion criteria were defined.

McCrindle (2013) performed a secondary analysis of an RCT conducted at 6 centers in patients who underwent the Fontan procedure. Exclusion criteria were a recognized indication for long-term anticoagulation; patient characteristics increasing the risk of hemorrhagic complications; known contraindication for heparin, warfarin, or acetylsalicylic acid (ASA); and the inability to supervise therapy because of social or geographic circumstances.

Potter (2013) conducted a cohort study in children who underwent the Fontan procedure and were identified from the New England Fontan registry, which included all patients who lived in the New England area, were born before 1985, and had Fontan surgery between April 1973 and July 1991 at Boston Children's Hospital. Exclusion criteria were surgery limited to cavopulmonary shunts, including those patients with an interrupted inferior vena cava and azygos extension to asuperior vena cava who underwent a bidirectional cavopulmonary shunt, and patients with early post-operative death (< 30 days) or a thromboembolic event within the first 14 days of Fontan surgery.

Table 2. Characteristics of included studies

_{*For further details, see risk of bias table in the appendix
¹Only overall age was reported
²Only overall sex was reported}

_{³Only overall sample size was reported
⁴Ischemic stroke is described as a separate outcome in this summary of the literature}

_{⁵Derived from Monagle, 2011 (original article)}

Results

1.Thrombotic complications

1.1. Studies with follow-up ≤ 1 year

Aspirin versus warfarin

Ankola (2021) reported the outcome thrombotic complications for the comparison of aspirin and warfarin with a follow-up of 30 days. The risk of thrombotic complications was higher in patients receiving aspirin (5/103; 4.9%) compared to those receiving warfarin (2/52; 3.8%), with RR = 1.26 (95% CI 0.25 to 6.29). The difference between the two groups is considered clinically relevant.

Aspirin versus enoxaparin

Ankola (2021) compared aspirin and enoxaparin for the outcome thrombotic complications with a follow-up of 30 days. Patients treated with aspirin had a lower risk of thrombotic complications (5/103; 4.9%) compared to those treated with enoxaparin (1/13; 7.7%), with RR = 0.63 (95% CI 0.08 to 4.99). The difference between the two groups is considered clinically relevant.

Warfarin versus enoxaparin

Ankola (2021) compared aspirin and enoxaparin for the outcome thrombotic complications with a follow-up of 30 days. Treatment with warfarin was associated with a lower risk of thrombotic complications (2/52; 3.8%) compared to treatment with enoxaparin (1/13; 7.7%), with RR = 0.50 (95% CI 0.05 to 5.10). The difference between the two groups is considered clinically relevant.

Aspirin versus rivaroxaban

McCrindle (2021) compared aspirin and rivaroxaban for the outcome thrombotic complications, with a follow-up period of 12 months. Patients treated with aspirin had a higher risk of thrombotic complications (2/34; 5.9%) compared to patients treated with rivaroxaban (1/64; 1.6%), with RR = 3.76 (95% CI 0.35 to 40.04). The difference between the two groups is considered clinically relevant.

1.2. Studies with follow-up > 1 year

Aspirin versus warfarin

Seven studies (Al-Jazairi, 2019; Egbe, 2017; Iyengar, 2016; McCrindle, 2013; Pessotti, 2014; Potter, 2013; Seipelt, 2002) compared aspirin and warfarin for the outcome thrombotic complications. Treatment with aspirin was associated with a higher risk of thrombotic complications compared to treatment with warfarin, with RR = 1.42 (95% CI 0.55 to 3.67; figure 2). The difference between the two groups is considered clinically relevant.

Figure 2. Forest plot of aspirin versus warfarin for the outcome thrombotic complications

Al-Jazairi (2019) reported an incidence rate of 7.82 events per 1000 patient years for warfarin, versus 5.83 per 1000 patient years for aspirin, giving an incidence rate ratio of 1.34 (95% CI 0.68 to 2.60). The difference in thrombotic complications between the groups is clinically relevant.

Iyengar (2016) reported a hazard ratio (adjusted for factors including among others sex, age at Fontan and number of pre-Fontan procedures) of 2.3 (95% CI 0.7 to 7.4) for the comparison of warfarin and aspirin, meaning that warfarin is associated with a higher risk of thrombotic complications compared to aspirin. This is a clinically relevant difference.

Seipelt (2002) reported a higher incidence rate of thrombotic complications for patients receiving aspirin (1.6 events per 100 patient years) compared to those receiving warfarin (1.1 per 100 patient years), giving an incidence rate ratio of 1.45. The difference between the groups is clinically relevant.

DOACs versus antiplatelets

Kawamatsu (2021) compared DOACs and antiplatelets for the outcome thrombotic complications, with a mean follow-up of 95 ± 64 months. Treatment with DOACs was associated with a lower risk of thrombotic complications (0/36; 0%) compared to treatment with antiplatelets (3/43; 7.0%), with RR = 0.17 (95% CI 0.01 to 3.18). The difference between the two groups is considered clinically relevant.

The incidence rate of thrombotic complications was 0% per patient year for DOACs, compared to 0.68% per patient year for antiplatelets, again reflecting a clinically relevant difference between the groups. The analysis was adjusted for duration after the Fontan procedure.

DOACs versus VKAs

Kawamatsu (2021) compared DOACs and VKAs for the outcome thrombotic complications, with a mean follow-up of 95 ± 64 months. Treatment with DOACs was associated with a lower risk of thrombotic complications (0/36; 0%) compared to treatment with VKAs (5/41; 12.2%), with RR = 0.10 (95% CI 0.01 to 1.80). The difference between the two groups is considered clinically relevant.

The incidence rate of thrombosis was 0% per patient year for DOACs, compared to 1.7% per patient year for VKAs. This is a clinically relevant difference between the groups.

DOACs versus combination of an antiplatelet and anticoagulant

Kawamatsu (2021) compared DOACs and a combination of an antiplatelet and anticoagulant for the outcome thrombotic complications, with a mean follow-up of 95 ± 64 months. No thrombotic complications occurred in either of the two groups. 

Antiplatelets versus VKAs

Kawamatsu (2021) compared antiplatelets and VKAs for the outcome thrombotic complications, with a mean follow-up of 95 ± 64 months. Treatment with antiplatelets was associated with a lower risk of thrombotic complications (3/43; 7.0%) compared to treatment with VKAs (5/41; 12.2%), with RR = 0.57 (95% CI 0.15 to 2.24). The difference between the two groups is considered clinically relevant.
The incidence rate of thrombotic complications was 0.7% per patient year for antiplatelets, compared to 1.7% for VKAs. This is a clinically relevant difference between the groups.

Antiplatelets versus combination of an antiplatelet and anticoagulant

Kawamatsu (2021) compared antiplatelets and VKAs for the outcome thrombotic complications, with a mean follow-up of 95 ± 64 months. Treatment with antiplatelets was associated with a higher risk of thrombotic complications (3/43; 7.0%) compared to treatment with a combination of an antiplatelet and anticoagulant (0/14; 0%), with RR = 2.39 (95% CI 0.13 to 43.57). The difference between the two groups is considered clinically relevant.

The incidence rate of thrombotic complications was 0.7% per patient year for antiplatelets compared to 0% per patient year for the combination of an antiplatelet and anticoagulant, which is a clinically relevant difference.

VKAs versus combination of an antiplatelet and anticoagulant

Kawamatsu (2021) compared VKAs and a combination of an antiplatelet and anticoagulant for the outcome thrombotic complications, with a mean follow-up of 95 ± 64 months. Treatment with VKAs was associated with a higher risk of thrombotic complications (5/41; 12.2%) compared to treatment with a combination of an antiplatelet and anticoagulant (0/14; 0%), with RR = 3.93 (95% CI 0.23 to 66.86). The difference between the two groups is considered clinically relevant.

The incidence rate of thrombotic complications was 1.7% per patient year for VKAs, compared to 0% per patient year for the combination of an antiplatelet and anticoagulant. This is a clinically relevant difference.

2. Hemorrhage

2.1. Studies with follow-up ≤ 1 year

2.1.1. Major bleeding

Aspirin versus rivaroxaban

McCrindle (2021) reported that no major bleeding events occurred in the aspirin group, while one major bleeding event occurred in the rivaroxaban group (1/64; 1.6%) during a 12-month follow-up period, with RR = 0.62 (95% CI 0.03 to 14.80). The difference between the two groups is clinically relevant.

2.1.2. Clinically relevant nonmajor bleeding

Aspirin versus rivaroxaban

McCrindle (2021) reported a higher proportion of clinically relevant nonmajor bleedings (3/34; 9%) in the aspirin group compared to the rivaroxaban group (4/64; 6%) during a 12-month follow-up period, with RR = 1.41 (95% CI 0.34 to 5.95). The difference between the groups is considered clinically relevant.

2.1.3. Trivial bleeding events

Aspirin versus rivaroxaban

McCrindle (2021) reported that the proportion of trivial bleedings was higher in the group of patients treated with aspirin (12/34; 35%) compared to the patients treated with rivaroxaban (21/64; 33%) during a 12-month follow-up period, with RR = 1.08 (95% CI 0.61 to 1.91). The difference between the groups is not considered clinically relevant.

2.2. Studies with follow-up > 1 year 

2.2.1. Major bleeding

Aspirin versus warfarin

Al-Jazairi (2019) compared aspirin and warfarin for the outcome major bleeding. The risk of major bleeding was lower in patients treated with aspirin (8/18; 44.4%) compared to patients treated with warfarin (9/18; 50%), with RR = 0.89 (95% CI 0.44 to 1.78). The difference is not considered clinically relevant.

Additionally, Al-Jazairi (2019) reported an incidence rate of 3.70 events per 1000 patient years for warfarin, versus 2.91 per 1000 patient years for aspirin, giving an incidence rate ratio of 1.34 (95% CI 0.68 to 2.60). Warfarin was associated with a higher incidence rate of major bleeding events compared to aspirin, the difference between the groups is clinically relevant.

McCrindle (2013) compared aspirin and warfarin for the outcome major bleeding. In both groups one major bleeding event occurred, the risk was 2% (1/57) in the aspirin group and 2% (1/54) in the warfarin group, with RR = 0.95 (95% CI 0.06 to 14.77). The difference is not clinically relevant.

The pooled relative risk for the comparison of aspirin and warfarin is 0.89 (95% CI 0.74 to 1.08), meaning that aspirin is associated with a lower risk of major bleeding events compared to warfarin (Figure 3). The difference is not considered to be clinically relevant.

Figure 3. Forest plot of aspirin versus warfarin for the outcome major bleeding

DOACs versus antiplatelets

Kawamatsu (2021) compared DOACs and antiplatelets for the outcome major bleeding, with a mean follow-up of 95 ± 64 months. The proportion of major bleeding events was lower in the group of patients treated with DOACs (1/36; 2.8%) compared to the group treated with antiplatelets (3/43; 7.0%), with RR = 0.40 (95% CI 0.04 to 3.66). This represents a clinically relevant difference.

The incidence rate of major bleeding was 0.6% per patient year for DOACs, compared to 5.1% per patient year for antiplatelets. This is a clinically relevant difference. The analyses was adjusted for duration after the Fontan procedure.

DOACs versus VKAs

Kawamatsu (2021) compared DOACs and VKAs for the outcome major bleeding, with a mean follow-up of 95 ± 64 months. Treatment with DOACs was associated with a lower risk of major bleeding events (1/36; 2.8%) compared to treatment with VKAs (6/41; 14.6%), with RR = 0.19 (95% CI 0.02 to 1.50). The difference between the groups is clinically relevant.

The incidence rate of major bleeding was 0.6% per patient year for DOACs, compared to 2.0% per patient year for VKAs, which is a clinically relevant difference.

DOACs versus combination of an antiplatelet and an anticoagulant

Kawamatsu (2021) compared DOACs and a combination of an antiplatelet and an anticoagulant for the outcome major bleeding, with a mean follow-up of 95 ± 64 months. Treatment with DOACs was associated with a lower risk of major bleeding events (1/36; 2.8%) compared to treatment with a combination of an antiplatelet and an anticoagulant (8/14; 57.1%), with RR = 0.05 (95% CI 0.01 to 0.35). The difference between the groups is clinically relevant.

The incidence rate of major bleeding was 0.6% per patient year for DOACs, compared to 5.1% for the combination of an antiplatelet and an anticoagulant. The difference is clinically relevant.

Antiplatelets versus VKAs

Kawamatsu (2021) compared antiplatelets and VKAs for the outcome major bleeding, with a mean follow-up of 95 ± 64 months. Treatment with antiplatelets was associated with a lower risk of major bleeding events (3/43; 7.0%) compared to treatment with VKAs (6/41; 14.6%), with RR = 0.48 (95% CI 0.13 to 1.78). The difference between the groups is clinically relevant.

The incidence rate of major bleeding was 0.7% per patient year for antiplatelets, compared to 2.0% per patient year for VKAs, which is a clinically relevant difference.

Antiplatelets versus combination of an antiplatelet and an anticoagulant

Kawamatsu (2021) compared antiplatelets and a combination of an antiplatelet and an anticoagulant for the outcome major bleeding, with a mean follow-up of 95 ± 64 months. Treatment with antiplatelets was associated with a lower risk of major bleeding events (3/43; 7.0%) compared to treatment with a combination of an antiplatelet and an anticoagulant (8/14; 57.1%), with RR = 0.12 (95% CI 0.04 to 0.40). This represents a clinically relevant difference between the groups.

The incidence rate of major bleeding was 0.7% per patient year for antiplatelets, compared to 5.1% per patient year for the combination of an antiplatelet and an anticoagulant, again reflecting a clinically relevant difference.

VKAs versus combination of an antiplatelet and an anticoagulant

Kawamatsu (2021) compared VKAs and a combination of an antiplatelet and an anticoagulant for the outcome major bleeding, with a mean follow-up of 95 ± 64 months. Treatment with VKAs was associated with a lower risk of major bleeding events (6/41; 14.6%) compared to treatment with a combination of an antiplatelet and an anticoagulant (8/14; 57.1%), with RR = 0.26 (95% CI 0.11 to 0.61). This is a clinically relevant difference.

The incidence rate of major bleeding was 2.0% per patient year for VKAs, compared to 5.1% per patient year for the combination of an antiplatelet and an anticoagulant. The difference is clinically relevant.

2.2.2. Minor bleeding

Aspirin versus warfarin

Al-Jazairi (2019) compared aspirin and warfarin for the outcome clinically relevant nonmajor bleeding. Treatment with aspirin was associated with a lower risk of minor bleeding events (25/66; 37.9%) compared to treatment with warfarin (41/66; 62.1%), with RR = 0.61 (95% CI 0.42 to 0.88). The difference between the groups is clinically relevant.

McCrindle (2013) compared aspirin and warfarin for the outcome minor bleeding with a follow-up period of 2.5 years. Treatment with aspirin was associated with a lower risk of minor bleeding events (8/57; 14%) compared to treatment with warfarin (18/54; 33%), with RR = 0.42 (95% CI 0.20 to 0.89). The difference between the groups is clinically relevant.

The pooled relative risk for the comparison of aspirin and warfarin is 0.57 (95% CI 0.09 to 3.61), meaning that aspirin is associated with a lower risk of minor bleeding events compared to warfarin (Figure 4). The difference is considered to be clinically relevant.

Figure 4. Forest plot of aspirin versus warfarin for the outcome minor bleeding

DOACs versus antiplatelets

Kawamatsu (2021) compared DOACs and antiplatelets for the outcome minor bleeding, with a mean follow-up of 95 ± 64 months. Treatment with DOACs was associated with a higher risk of minor bleeding events (1/36; 2.8%) compared to treatment with antiplatelets (1/43; 2.3%), with RR = 1.19 (95% CI 0.08 to 18.43). The difference between the groups is not considered clinically relevant.

DOACs versus VKAs

Kawamatsu (2021) compared DOACs and VKAs for the outcome minor bleeding, with a mean follow-up of 95 ± 64 months. Treatment with DOACs was associated with a lower risk of minor bleeding events (1/36; 2.8%) compared to treatment with VKAs (3/41; 7.3%), with RR = 0.38 (95% CI 0.04 to 3.49). The difference between the groups is clinically relevant.

DOACs versus combination of an antiplatelet and an anticoagulant

Kawamatsu (2021) compared DOACs and a combination of an antiplatelet and an anticoagulant for the outcome minor bleeding, with a mean follow-up of 95 ± 64 months. Treatment with DOACs was associated with a higher risk of minor bleeding events (1/36; 2.8%) compared to treatment with a combination of an antiplatelet and anticoagulant (0/14; 0%), with RR = 1.22 (95% CI 0.05 to 28.21). This is not a clinically relevant difference between the groups.

Antiplatelets versus VKAs

Kawamatsu (2021) compared antiplatelets and VKAs for the outcome minor bleeding, with a mean follow-up of 95 ± 64 months. Treatment with antiplatelets was associated with a lower risk of minor bleeding events (1/43; 2.3%) compared to treatment with VKAs (3/41; 7.3%), with RR = 0.32 (95% CI 0.03 to 2.93). The difference between the groups is clinically relevant.

Antiplatelets versus combination of an antiplatelet and anticoagulant

Kawamatsu (2021) compared antiplatelets and a combination of an antiplatelet and anticoagulant for the outcome minor bleeding, with a mean follow-up of 95 ± 64 months. Treatment with antiplatelets was associated with a higher risk of minor bleeding events (1/43; 2.3%) compared to treatment with a combination of a antiplatelets and anticoagulants (0/14; 0%), with RR = 1.02 (95% CI 0.04 to 23.78). The difference is not clinically relevant.

VKAs versus combination of an antiplatelet and anticoagulant

Kawamatsu (2021) compared VKAs and a combination of an antiplatelet and anticoagulant for the outcome minor bleeding, with a mean follow-up of 95 ± 64 months. Treatment with VKAs was associated with a higher risk of minor bleeding events (3/41; 7.3%) compared to treatment with a combination of an antiplatelet and anticoagulant (0/14; 0%), with RR = 2.50 (95% CI 0.14 to 45.62). This is considered a clinically relevant difference.

2.2.3. Clinically relevant nonmajor bleeding

Aspirin versus warfarin

Al-Jazairi (2019) compared aspirin and warfarin for the outcome clinically relevant nonmajor bleeding. Treatment with aspirin was associated with a higher risk of clinically relevant nonmajor bleeding events (5/16; 31.3%) compared to treatment with warfarin (11/16; 68.8%), with RR = 0.45 (95% CI 0.20 to 1.01). The difference between the groups is clinically relevant.

3. Mortality

3.1. Studies with follow-up ≤ 1 year

No evidence was found for the outcome mortality.

3.2. Studies with follow-up > 1 year

Aspirin versus warfarin

Pessotti (2014) compared aspirin and warfarin for the outcome mortality, with a follow-up period of 24 months and showed that no patients died in the aspirin group, compared to two in the warfarin group (2/15; 13.3%), with RR = 0.20 (95% CI 0.01 to 3.85). The difference between the groups is considered clinically relevant.

McCrindle (2013) compared aspirin and warfarin for the outcome mortality, with a follow-up period of 2.5 years and showed that one patient died in the aspirin group (1/57; 2%), compared to no patients in the warfarin group (0%), with RR = 2.84 (95% CI 0.12 to 68.36). The difference between the groups is clinically relevant.

4. Ischemic stroke

4.1. Studies with follow-up ≤ 1 year

Aspirin versus rivaroxaban

McCrindle (2021) reported that ischemic stroke occurred in one patient in the aspirin group (1/34; 3%), while no patients in the rivaroxaban group had an ischemic stroke, RR = 5.6 (95% CI 0.23 to 133.19). The difference between the groups is considered to be clinically relevant.

4.2. Studies with follow-up > 1 year

No evidence was found for the outcome ischemic stroke.

5. Quality of life (QOL)

No evidence was found for the outcome quality of life (QOL).

6. Adverse events

6.1. Studies with follow-up ≤ 1 year

Aspirin versus rivaroxaban

McCrindle (2021) reported that the proportion of patients experiencing at least one adverse effect during the study period was similar in both groups (29/34; 85% in patients receiving aspirin and 55/64; 86% in patients treated with rivaroxaban), with RR = 0.99 (95% CI 0.84 to 1.18).

6.2. Studies with follow-up > 1 year

No evidence was found for the outcome adverse events.

A systematic review of the literature was performed to answer the following question(s):

What are the favorable and unfavorable effects of the use of thromboprophylaxis A compared with the use of thromboprophylaxis B for children with an indication for the Fontan procedure?

Table 1. PICO

Relevant outcome measures

The guideline panel considered thrombotic complications (including pulmonary embolism) as a critical outcome measure for decision making; and the other outcomes as important outcome measures for decision making.

A priori, the guideline panel did not define the outcome measures listed above but used the definitions used in the studies.

The guideline panel defined a difference of 25% (0.8 ≥ RR ≥ 1.25) as minimal clinically (patient) important difference.

Search and select (Methods)

A systematic literature search was performed by a medical information specialist using the following bibliographic databases: Embase.com and Ovid/Medline. Both databases were searched from 2014 (publication of former Dutch guideline) to 9 July 2024 for systematic reviews, RCTs and observational studies. Systematic searches were completed using a combination of controlled vocabulary/subject headings (e.g., Emtree-terms, MeSH) wherever they were available and natural language keywords. The overall search strategy was derived from two primary search concepts: (1) Fontan circulation; (2) thromboprophylaxis/anticoagulants. Duplicates were removed using EndNote software. After deduplication a total of 479 records were imported for title/abstract screening.

Studies were selected based on the following criteria:

• Study design: systematic reviews in which searches were performed in at least two databases, with a detailed search strategy, risk of bias assessment and results of individual studies available, randomized controlled trials, or (observational) comparative studies;
• Full-text English or Dutch language publication;
• Studies according to the PICO; and
• Studies including at least twenty participants (ten per group).

Initially, 24 studies were selected based on title and abstract screening. After reading the full text, 17 studies were excluded (see the exclusion table under the tab ‘Evidence tabellen’), and seven studies were included.

In addition to the systematic literature search, we also screened the title and abstract of four additional studies that were included in the conclusions of the module on Fontan in the TRAMPOLINE protocol (former Dutch guideline). One study (Khairy, 2008) did not fit the PICO (no comparison between thromboprophylactic strategies was made). After reading the full text, three studies were included in this summary of literature (Seipelt, 2002; McCrindle, 2013; Potter, 2013).