Application of Estimated Glomerular Filtration Rate Decline Rate to Optimise Timing of Pre-Emptive Arteriovenous Fistula Creation for Chronic Kidney Disease Patients.

Abstract

Aim: Establishing a functional pre-emptive arteriovenous fistula (pAVF) before the commencement of haemodialysis requires careful patient selection and precise timing. Our study aimed to create an eGFR (estimated glomerular filtration rate) slope-based prediction model to estimate the time before the onset of haemodialysis for individual patients.

Methods: In this prognostic prediction study, we retrospectively analysed past eGFR data from chronic kidney disease patients who underwent pAVF surgery at Naresuan University Hospital and Maharaj Nakorn Chiangmai Hospital in Thailand.

Results: Seventy-one patients who underwent pAVF creation between January 1st, 2016, and December 31st, 2022, were included. 75% achieved successful AVF cannulation for their initial haemodialysis, with 5% not needing dialysis. Past eGFR decline rate (eGFRr), eGFR at referral for AVF (eGFRbase), age ≥ 80 years systolic blood pressure, body mass index, and serum calcium levels have been identified as significant factors in estimating the time remaining before the initiation of haemodialysis (Tr). The model estimates the probability of successful pAVF use versus the need for a catheter at the first haemodialysis session. The model's discriminative performance, evaluated via area under the receiver operating characteristic curve (AuROC), yielded a value of 0.96 (95% CI: 0.92, 1.00) and reliable results as confirmed by the calibration plot.

Conclusion: Our prediction model helps estimate Tr for individual patients, enabling timely referral for creation of a pAVF. Implementation in clinical practice has the potential to enhance the utilisation of pAVF and reduce the risk of catheter-dependent haemodialysis. However, a prospective validation study is required to confirm the accuracy of the model before its clinical application.