Optimal long-term vascular access approach in chronic kidney disease patients for hemodialysis

Arteriovenous fistula (AVF) is the best option of vascular access for maintenance hemodialysis (HD), owing to its prolonged patency, optimal blood flow, few complications, and low costs to maintain. Promoting the use rate of AVF is the most optimal approach for the vascular access in HD patients, bringing many benefits to them and minimizing treatment costs. However, the rates of AVF maturation success and patency are still low, expanding the rate of AVF ultization in HD patients remains challenging.
Well vessel preparation, pre-operative evaluation and determining suitable time for AVF creation is the first step to improve the outcome of AVF maturation. AVF creation is personalized for each patient based on initial dialysis time, helping AVF mature in time for dialysis and without wasting its usage time. However, determining when to start dialysis in each patient is not simple. Herein we introduced a machine learning model to predict the initial renal replacement therapy in chronic kidney disease (CKD) patients with advanced stage. Important variables such as demographics, comorbidities, medications, blood and urine tests from 5279 advanced CKD patients were collected. Several machine learning models were developed and compared, of which the gradient boosting classifier model showed better performance. This model reaches an AUC of 0.91, with a sensitivity and specificity of 0.84 and 0.82, respectively. Thus, we successfully developed the machine learning model for accurately predicting the initial RRT in advanced CKD patients. The future work consists of applying the model in a clinical trial and further in clinical practice to aid in estimating the time of initial RRT, thereby making a pre-RRT treatment plan and improving the patient outcome.
The next steps are to identify and control its novel risk factors that are of great importance for improving AVF maturation and patency rates. Laxatives are frequently used to treat constipation, but they are also known to be an independent factor in adverse cardiovascular events. Since AVF is part of the cardiovascular system, its maturation might be affected by the same risk factors for cardiovascular disease. Therefore, contact laxative use might be a new risk factor for AVF maturation failure. Our second study investigated the relationship between contact laxative usage and AVF maturation outcome in HD patients. 480 contact laxative users and 472 non-users who had received initial AVF creation participated in our multi-center cohort. They were monitered until the results of AVF maturation were verified. The risk of AVF maturation failure brought on by laxatives was assessed using multivariable logistic regression models. In this study, we discovered that patients who used contact laxatives had a considerably higher risk of AVF maturation failure than non-users (adjusted odds ratio, 1.64; p=0.003). Notably, the risk of AVF maturation failure increased when increasing their average daily doses and cumulative treatment days. In other words, our research revealed a definite dose- and duration-dependent association between contact laxative usage and a higher risk of AVF maturation failure. Laxatives should therefore be prescribed with caution in this population. Additional studies are required to verify these findings and investigate the potential mechanisms.