Discussion

Obtaining the best access for a planned cardiac catheterisation is the primary key point for success. In the absence of femoral access, transjugular and transhepatic routes stand out as reasonable alternates for atrial septal defect closure. In adults, the internal jugular vein is a more classical entry point for catheterisation but addressing a centrally located atrial septal defect at a perpendicular angle may be encountered with some technical difficulties especially in patients with dilated right cavities. The puncture of the subclavian vein has been reported in complex catheterisations but can be easily complicated with pneumothorax and bleeding in a noncompressible site. The straight forward access to the atrial cavities via transhepatic access and the ease to introduce large delivery systems without serious risk of vascular injury motivated us to consider this route for defect closure in our patient with significant morbidities.^{Reference Shim, Lloyd and Beekman5,Reference Ebeid6} The occluder alignment with the plane of the septum avoided caudal traction on the delivery cable. This feature was also highlighted in closing the large atrial defects with deficient rims, avoiding complex manoeuvers, and prolapse of any part of the occluder.^{Reference Ebeid6}

Some congenital interventionists have taken over the technique in some institutions and reported favourable experiences with variable success rates.^{Reference Shim, Lloyd, Cho, Moorehead and Beekman1,Reference Emmel, Sreeram, Pillekamp, Boehm and Brockmeier7,Reference McLeod, Houston, Richens and Wilson8} McLeod et al. even reported obtaining the hepatic access blind without the assistance of ultrasound and stated that as operator experience increases, the need for ultrasound guidance is reduced.^{Reference McLeod, Houston, Richens and Wilson8} However, we believe that the presence of an experienced interventional radiologist is unvoidable during first attempts. The knowledge and the skills to perform this approach should be acquired by the cardiac interventionalist, to be used when deemed necessary.^{Reference Ebeid6} Appropriate ultrasound evaluation of the liver anatomy and location is mandatory in grown-up congenital heart patients to identify the suitable hepatic vein for access. Ultrasound guidance is also needed during needle entry and sheath placement, in adjunction to fluoroscopy to optimise procedural safety and outcomes.

The transhepatic approach can be easily complicated with liver hematoma and intraperitoneal bleeding although the absolute risk is still not certain.^{Reference McLeod, Houston, Richens and Wilson8} The liver can store up to 10% of the total blood volume at any given time, and this volume can significantly increase in cases of elevated central venous pressure.^{Reference Guyton and Hall9} The antithrombotic state was carefully evaluated before the intervention. The medication was uninterrupted since our patient was estimated at a high risk of thromboembolism and heparin was added for atrial manipulations and bidirectional shunting. Access closure and heparin reversal before sheath withdrawal remain until now a matter for debate.^{Reference Ebeid6} Some authors reported achieving haemostasis in children only by manual external compression of skin at the puncture site with no significant complications.^{Reference Emmel, Sreeram, Pillekamp, Boehm and Brockmeier7,Reference McLeod, Houston, Richens and Wilson8} Others consider the closure of the tract only in patients who are on aspirin or in those with elevated right atrial pressure.^{Reference Ebeid6} Our patient had a basal right atrial pressure of 12 mmHg, required the use of a large sheath, and his procedure unexpectedly elapsed fast. Active clotting time had to be kept below 200 seconds at the end of the procedure before Gelfoam plugs were used to achieve fast and efficient hepatic vein haemostasis and allow early mobilisation. Although it does not cause artefacts in MRI, the gelfoam radiolucency was considered as an inconvenience for controlled and precise parenchymal access embolisation. Only detachable coils and vascular plugs were reported as haemostatic alternatives in similar clinical situations.^{Reference Alkhouli, Campsey, Higgins, Badhwar, Diab and Sengupta3,Reference Jolly, Dhar and Amin4,Reference Ebeid6} However, the risk of infectious complications has to be considered when using vascular plugs. Dollinger et al. reported a case of focal liver abscess adjacent to a vascular plug.^{Reference Dollinger, Goessmann, Mueller-Wille, Wohlgemuth, Stroszczynski and Heiss10} With this in mind, Uller et al. state that the gelfoam application does not increase the risk of local inflammation in comparison to permanent embolic agents offering a major advantage especially in patients with diminished immunity.^{Reference Uller, Müller-Wille and Grothues11}