Back

Planning optimal sequence for arteriovenous fistula creation in hemodialysis patients, using quality-of-life adjusted access patency

Kumar, M.; Usulumarty, D.; Brahmbhatt, R.; Rangaraj, N.; Noronha, S.; Billa, V.

2023-04-06 health informatics
10.1101/2023.04.06.23288219 medRxiv
Show abstract

ObjectiveThe study proposes a decision strategy for the choice of arm for placement of an arteriovenous fistula (AVF) in patients with kidney failure who undergo hemodialysis (HD). The possible surgical sites for placement of the AVF are located in the arm and at two locations in each arm, the wrist (distal AVF) and the elbow (proximal AVF). Recommendations are made for the location of the first AVF and subsequent ones, as needed. MethodA retrospective analysis of AVFs created between 2015 and 2022 in patients at five HD centers was performed. The study uses qualityadjusted survival, parametric, and nonparametric survival methods. We conducted two surveys to assess the HD patients quality of life. In addition, the Cox proportional hazard model is used to suggest a patient-specific strategy. The survival functions are compared using the log-rank and Tarone-Ware tests. ResultsThe results of the multivariate analysis showed that placing AVFs on the non-dominant arm leads to superior patient quality of life compared to the dominant arm. The quality-adjusted survival was also found to be better when AVFs were located on the non-dominant arm. Considering both these aspects and also clinical constraints on the sequence, the optimal sequence is found to be non-dominant distal followed by non-dominant proximal locations, followed by a similar sequence on the dominant hand if required. ConclusionThe study identifies criteria for data-driven decision-making as to good locations for AVFs for Hemodialysis and supports clinical experience and practice in this matter.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.