Optimisation of Radiocephalic Arteriovenous Fistula

Each year, an estimated 5,000-6,000 patients require haemodialysis across United Kingdom (UK). This is only possible with a functional, accessible and permanent arteriovenous fistula (AVF). The current guidelines recommend, radiocephalic AVF (RACVF) to be the primary choice of fistula. Prior to the main objectives of this thesis, a UK wide (clinician based) survey was conducted to ascertain the scope of practice and Functional Maturation (FM) of RCAVF in UK. The survey demonstrated that 40% of surgeons prefer not to create RACVF (due to various factors) despite these current recommendations. In addition, current FM in the UK is around 60%, a finding which corresponds with a systematic review of the literature (comprised of exclusively UK studies). This rate of FM is 12% less than that reported internationally (72%). Therefore, it was hypothesised (null hypothesis) that FM of RCAVF could not be enhanced by optimisation of preoperative evaluation, patient selection (demographics), comorbidities, anatomical factors, operative skills and postoperative care and follow up. A methodological & prospective study, with series of retrospective analytics on n=350 RCAVF over six years from creation to their FM or failure was conducted. The traditional statistical means was also complemented by the first feasibility and applicability study, using artificial intelligence (AI) and deep learning (DL) for pattern recognition, modelling and decision tree in RCAVF stratification and formation. The outcome suggests that the traditional clinical examination for patient selection, should be complemented by the use of preoperative duplex sonography through a multidisciplinary team approach. Patient demographics and comorbidities should not be used as an exclusion criterion for RCAVF formation once the minimum diameter of cephalic vein (>1.8mm) and radial artery (2mm) has been met. Arteriovenous Ratio (AVR) of 0.8 to 1.1 creates minimal alteration in pressure and velocity of RCAVF, limiting early failure. Almost 50% of non-functioning RCAVF through early surveillance could be salvaged by well-designed algorithmic salvage (open & endovascular) techniques. Application of all these outcomes could result in a FM of 80% in clinical practice which is a major improvement on the current rate of 60%. The future role of AI and its implementation with vast dataset could enhance stratification and FM. The findings of cephalic vein and radial artery diameter from this thesis has now influenced "The National Kidney Foundation's Kidney Disease Outcomes Quality Initiative (KDOQI)" and "European society of Vascular Surgery Guidelines". Finally, there is an international study (generated from this thesis) to gather the scope of practice (worldwide) and establish consensus.