Endovascular intervention in the maintenance and rescue of paediatric arteriovenous fistulae for hemodialysis

Arteriovenous fistulae (AVF) are recognised as the gold standard type of vascular access for hemodialysis (HD) in adults [1, 2]. A fistula-first policy for paediatric hemodialysis has been advocated since 2005 [3], yet international registries show that the majority of children are maintained on central venous catheters (CVC) [4, 5]. Hemodialysis via a functional AVF provides superior clinical outcomes compared to hemodialysis using a CVC, including quality of dialysis, lower infection rates and reduced hospitalisation [6]. However, AVFs also develop complications such as stenosis and thrombosis, requiring intervention to extend their functional patency. Fistula surveillance programmes, as recommended by the Kidney Disease Outcomes Quality Initative (KDOQI) guidelines [7], frequently identify AVF-related complications which can be managed by endovascular and open surgical methods. Perhaps due to the low volume of paediatric AVFs in use [8, 9], little has been written regarding the efficacy of endovascular intervention in children with AVF complications.

This case series aims to determine the outcome of an ‘endovascular-first’ approach to managing malfunction in paediatric AVFs for haemodialysis. It is the largest series to date.

A retrospective review of endovascular intervention for paediatric AVFs from 2005 to 2016 years was undertaken. Data were collected from two paediatric tertiary centres (Great Ormond Street Hospital and Evelina Children’s Hospital) London, UK. Medical and radiological case records were retrieved from the dialysis units’ computer databases and radiology archives. Participants were included if they underwent endovascular intervention for acute AVF thrombosis or clinically significant consequences of AVF stenoses affecting the quality of haemodialysis.

The primary end-point was immediate success of the endovascular procedure, defined as the ability to undergo effective hemodialysis through a patent AVF in the first 12 h following intervention. The secondary end-point was quality of dialysis at 3 months following endovascular intervention. Complications of angioplasty were also reported.

The angiographic images (‘fistulograms’) of the AVFs were independently reviewed by two consultant interventional radiologists, in order to determine lesion location and morphology. The severity of stenoses was calculated as a ratio of the maximum internal stenotic luminal diameter compared to the maximum internal luminal diameter in a normal segment of adjacent upstream vessel. The lesions’ anatomical location was described using the classification system outlined in Table 1.

The study reviewed all endovascular procedures undertaken between 2005 and 2016. Eighteen children were included in the study. All the interventions were performed as the first treatment modality for the AVF dysfunction.

Almost all children on a dialysis programme will be eligible for transplantation. Hence dialysis in childhood is seen as a ‘bridge’ until transplantation is possible. Currently, in the UK about 30% of children will experience haemodialysis before transplantation and the waiting time for first cadaveric renal transplant is approximately 1 year [10]. ‘Maintenance’ of dysfunctional AVFs is part of the natural history of managing vascular access and works in parallel with a surveillance programme. In adults, endovascular intervention is generally recognised as the preferred initial maintenance tool, but little is written about its role in children who have smaller AVFs and highly reactive vessels.

The indication for intervention was normally identified through a surveillance programme, unless the AVF had become acutely dysfunctional. Our current practice is for all children to undergo a 4-monthly surveillance assessment which facilitates early detection of AVF pathology. This involves an ‘ABCDE’ assessment including dialysis adequacy, volume flow measurement and clinical review (Table 4). In addition, we encourage an ‘open door’ policy for any concerns from the child/family or dialysis staff. This allows referral into a dedicated vascular access clinic with surgical, dialysis and ultrasound expertise. Families and referring healthcare staff are encouraged to screen for AVF-related complications using a clinical review checklist tool described in Table 5.

Our experience suggests that endovascular techniques are a valuable ‘first line’ tool for the maintenance of paediatric AVFs. Most cases can be performed as day case procedures with the child being prepared as an outpatient during haemodialysis and admitted from home on the day of the procedure for intervention under general anaesthesia. Immediately following intervention, the child should have a ‘test dialysis session’ and if successful be discharged under supervision. Rapid turnaround is important, as many children travel significant distances and minimising their school disruption is important. A high success rate can be anticipated (86% in this study) with excellent dialysis metrics at 3 months post procedure.

Most children will rotate through a ‘Dialysis – Transplant – Dialysis’ cycle before reaching adulthood. Hence, AVF maintenance is essential even in those with an imminent transplant as the functioning AVF can be subsequently used for phlebotomy following transplantation, by suitably experienced staff. Additionally, the AVF remains viable for future dialysis and the venous reserves are preserved for the future. It should also be noted that loss of AVF in the post-transplant period due to thrombosis is frequent [11]. Possible causes of AVF failure include hypovolaemia, pro-thrombotic immunosuppression and inexperienced phlebotomy. This should be guarded against with maintenance of a euvolemic state and training of phlebotomy staff in AVF management. Additionally, we recommend the prescription of 1 mg/kg oral aspirin for all children with an AVF.

Clinicians should be aware of any anticipated date for transplantation, which is helpful in planning a maintenance strategy for a malfunctioning AVF. In the UK, most living donation is performed within 3–6 months from initiating donor work-up, whilst the first deceased-donor transplant waiting time is approximately 1 year [10]. Hence, ‘nursing’ a functioning AVF with repeated endovascular interventions is preferable to abandoning it and electing for a CVC, especially if a transplant is on the horizon. Even modest improvements in luminal diameter and volume flows are worth achieving and often allow successful dialysis from a previously malfunctioning AVF. In our experience, endovascular intervention has a low risk profile (3%) and any resulting complications (e.g. extravasation, rupture or embolism) can usually be managed by additional endovascular techniques.

Failure to rescue a malfunctioning AVF using endovascular therapy (~ 14%) is most commonly seen in AVFs with multiple lesions. In such cases, surgical intervention may be a viable option in order to avoid the use of a CVC. In reality, this requires urgent coordination between the vascular access surgeon, anaesthetist and emergency theatres. The pathway of least resistance is often to defer to a CVC, but this should be resisted. The dialysis access nurse coordinator plays a pivotal role in prioritising such cases. When undertaking surgical revision, care should be taken in pre-operative planning of how to deliver dialysis post-operatively. Establishing new needling sites away from any areas of post-operative swelling and using single needle techniques may allow continued use of the AVF without introducing a CVC. If the child is self-needling on a home haemodialysis programme, it may be prudent to transfer back to in-centre dialysis with experienced dialysis staff until two permanent needling sites can be re-established. Compared with a surgical revision, minimally invasive interventions are associated with less pain, better cosmesis, shorter duration of hospitalisation and greater chance of avoiding a CVC.

One of the major advantages of an ‘endovascular-first’ approach to AVF malfunction is the avoidance of a CVC. CVCs have been shown to offer inferior dialysis adequacy, greater infection risk and higher hospitalisation rates compared to AVF-based haemodialysis [6] in children. Avoiding a CVC is also vital in preserving the central venous system for future AVF formation in children on chronic haemodialysis. However, currently available databases suggest this recommended practice is not routinely observed [4, 5]. In our experience, once a CVC is inserted there is often resistance from the child or parent to re-establishing dialysis using an AVF. Even the use of temporary dialysis catheters for short durations is not risk-free. These cause damage to central veins and increase risk for insertion-site and systemic infections. Temporary CVCs placed in the neck veins may negatively impact on future AVF formation in the upper limbs. When placed in the lower limbs, they may damage the femoral and iliac veins, making future renal transplantation more complex. Hence, AVF malfunction should be prioritised and follow an ‘emergency protocol’ for urgent assessment and rapid access to endovascular intervention where necessary.

The authors recognise that the study’s data is limited by the small population size and its retrospective nature, which did not allow sub-group analysis and measurement of long-term outcomes data.

The management of paediatric AVF dysfunction with an ‘endovascular-first’ approach provides an effective solution for maintaining and rescuing vascular access in children on haemodialysis, with few associated complications. Repeated endovascular interventions may be necessary to avoid CVCs, until renal transplantation is possible. We recommend the development of emergency protocols and checklists for managing malfunctioning AVFs.