Abstract Background and Aims Vascular access (VA) is vital for hemodialysis (HD) treatments and should allow to achieve the recommended dose of dialysis. Thrombosis is the primary cause of VA failure in grafts (AVG) and fistulas (AVF) and should be avoided as it can lead to significant patient morbidity and mortality. However, despite its importance, predictors of VA thrombosis are not completely established. Even if large randomized controlled trials are lacking to clearly identify the ideal surveillance strategies and benefits of surveillance, evidence suggests that when combined with preemptive interventions, prospective monitoring and surveillance and early detection of dysfunctional VA may reduce the thrombosis rate and improve its long-term patency. Several monitoring and surveillance techniques of VA can be used, from unexplained drop in dialysis adequacy (Kt/V and convective volume) to the noninvasive estimation of VA blood flow (Qa). The aim of this study was to assess whether these parameters can predict VA thrombosis. Method We retrospectively evaluated all episodes of VA thrombosis admitted in our Vascular Access Centre (VAC) between July 2019 and December 2022. Patient's demographics, VA characteristics and data from dialysis sessions were collected. Descriptive statistics were calculated and expressed as mean (standard deviation) or median (IQR) or count (%), as adequate. Values were compared using Mann-Whitney for independent samples and Pearson chi-squared test with 95% confidence intervals (CI). The predictors of VA thrombosis studied were Qa, Kt/V and convective volume (CV), in the case of the patients on hemodiafiltration. Qa was measured using the thermodilution method called blood temperature monitoring (BTM®), and Kt/V and CV with Online Clearance Monitoring (OCM®). In this study, we evaluated the ratio between the last measured value of Qa prior to the thrombosis episode and the average of the previous three months (Qa trend), and the ratio between the last measured value of Kt/V and CV before the thrombosis episode and the average of the previous month (Kt/V trend and CV trend, respectively) as predictors of VA thrombosis Results From July 2019 and December 2022, we treated, in our VAC, 386 VA thrombosis episodes in 226 patients. In our study population, mean age was 71.1 (± 13.2) years, 64% were male, 31% had diabetes. The dialysis vintage was 51 (69) months and the VA vintage was 42 (62) months. 70.2% of the VA thrombosis were on AVF. Prior to thrombosis the median of Qa obtained by BTM was 650 (495) mL/min, the median of kt/v was 1.43 (0.62) and the median of CV was 22.4 (6.3). The average of the previous values for Qa, Kt/V and CV were 658 (353) ml/min, 1.53 (0.35) and 23.5 (3.9) L. The median (IQR) of Qa trend, Kt/V trend and CV trend were -8,5 (33,3) %, -6,0 (28,2) % and -2,5 (17,0)%. We found no statistical significance for Qa trend and CV trend between AVF and AVG as predictors of VA thrombosis, but the reduction in Kt/V values was significantly more pronounced in AVF (-6.8%; Q1 -30.1% to Q3 0.7%) than in AVG (-3.8%; Q1 -18.0% to Q3 + 5.0%) (p = 0.007) as predictor of thrombosis. In 68.9% of VA thrombosis, we noticed a reduction in Kt/V value before the episode (73.8% in AVF and 57.5% in AVG, p = 0.002). We found no statistically significance in the reduction of Qa and CV values between AVF and AVG thrombosis episodes. Conclusion Our study showed that Kt/V trend was a better predictor of VA thrombosis than Qa trend and CV trend, especially in AVF.
Read full abstract