Frailty screening for determination of hemodialysis access placement

Abstract

ObjectiveChoosing the right hemodialysis vascular access for frail patients remains difficult because the patient's preferences and the likelihood of access function and survival must be considered. We hypothesize that patients identified before arteriovenous (AV) access as frail by the PRISMA-7 score may have worse outcomes, indicating that fistula creation may not be the most clinically beneficial option and it would be in the best interest of the patient to receive either AV graft (AVG) placement or dialysis through a percutaneous catheter. Our pilot study aims to determine whether an association exists between patient frailty as defined by PRISMA-7 and newly created AV fistula (AVF) and AVG access outcomes. MethodsThis was a single institutional prospective cohort study of patients undergoing new AVF or AVG intervention from April 2021 to May 2023. Patients were assessed using the PRISMA-7 frailty questionnaire before their AV access surgery. Patients were grouped by frailty score and score groups were examined for trends. Univariable analysis was performed for baseline differences between frail and nonfrail patients. Failure to achieve maturation, postoperative infection, and 180-day mortality difference was also investigated for frail vs nonfrail patients. Univariable analysis was performed for nonmaturation using standard comorbidities, arterial and venous diameters, and frailty. Multivariable binary logistic regression was performed for the outcome of nonmaturation using frailty as one of the variables in conjunction with the univariable risks associated with nonmaturation. ResultsA total of 40 patients undergoing new AV access placement were investigated, among whom 53% were designated as frail (PRISMA-7 score ≥3). When comparing the frail and nonfrail new AV access groups, the access (AVF and AVG combined) failed in 48% (10/21) of the frail patients, but only failed in 5% (1/19) of the nonfrail patients 1 (P = .012). When distinguishing between AV access types, AVF creations followed the overall trend with 60% of AVF access (9/15) sites in frail patients failing to mature when compared with nonfrail patients, who all had fistulas that matured to use (P = .049). Surgical site infection was absent in all frail patients and present in 5% of nonfrail patients (1/19). Both 30-day and 60-day readmission rates were higher in the frail group compared with the nonfrail group. There was 180-day mortality present in 5 of frail patients % (1/21) and absent in nonfrail patients. Multivariable analysis revealed that both frailty (adjusted odd ratio, 10.19; 95% confidence interval, 1.20-82.25); P = .033) and younger age (adjusted odd ratio, 0.953; 95% confidence interval, 0.923-0.983; P = .002) both had a significant association with nonmaturation. Power analysis revealed a power statistic of 0.898 indicating a probability of type 2 error of 10.02% with a P value of .002. Hosmer-Lemeshow goodness of fit for the logistic regression had 75% overall accuracy for the model. ConclusionsPatient frailty is significantly associated with an increased incidence of AV access failure to mature.