William D. Paulson

Accuracy of decrease in blood flow in predicting hemodialysis graft thrombosis.

We recently showed that a single low graft blood-flow measurement (Qa) does not accurately predict graft thrombosis. In this study, we prospectively determined whether percentage of decrease in Qa (DeltaQa) or adjustment of Qa for mean arterial pressure (Qa/MAP; Delta(Qa/MAP)) provides greater predictive accuracy than a single Qa. We monitored 83 grafts from 80 patients for thrombosis over periods up to 12 months. Qa (by ultrasound dilution) and MAP were measured monthly during the study. Receiver operating characteristic curves were used to determine whether Qa, DeltaQa, Qa/MAP, or Delta(Qa/MAP) provided the combination of high sensitivity (>80%) and low false-positive rate (FPR; <20%) needed for clinical use. This level of predictive accuracy requires an area under the curve (AUC) of approximately 0.90. We analyzed the four predictors by a number of criteria and found that all AUCs were less than 0.90 and adjustment for MAP reduced the AUC. In predicting thrombosis within 1 month, for example, AUCs for Qa and net DeltaQa (over 3 months) were 0.84 and 0.82, respectively, whereas AUCs for Qa/MAP and net Delta(Qa/MAP) were 0.78 and 0.75, respectively. At a sensitivity of 80%, FPRs for all predictors were at least 30%. Thus, a high sensitivity always required a high FPR. These results show that DeltaQa and adjustment for MAP are not more accurate than a single low Qa in predicting thrombosis. None of these predictors provide enough predictive accuracy to be the sole criterion for clinical decision making. A successful monitoring and intervention program will likely require the inclusion of other predictors that, together with Qa, may provide the needed accuracy.

Hemodynamic reproducibility during blood flow measurements of hemodialysis synthetic grafts.

We have previously shown that graft blood flow (Qa) has a poor accuracy in predicting graft thrombosis. In this study, we determined whether hemodynamic variation helps explain this poor predictive accuracy. We also determined whether standardized timing of Qa measurements, which is widely recommended, will promote measurement reproducibility. We analyzed variations in mean arterial pressure (MAP) in seven consecutive dialysis sessions for 51 patients and determined the influence of MAP on Qa (by ultrasound dilution). We used a pooled coefficient of variation (CV) to summarize MAP variation within individual patients (computed as +/-2 CVs). MAPs from the seven sessions varied widely, and most variation was present with the first MAPs at the beginning of the sessions. These first MAPs varied by +/-23%, whereas variation for the entire session was +/-28%. The influence of MAP on Qa was determined by measuring the two together during consecutive thirds of a single session. The percentage of change in MAP (DeltaMAP) and Qa (DeltaQa) from the first to middle or last thirds of the session varied over wide ranges: -37% to 86% and -43% to 78%, respectively. The DeltaQa versus DeltaMAP correlation was relatively strong for changes between the first and middle thirds (r = 0.666) and first and last thirds (r = 0.646) of the session (both P: < 0.01). We conclude that MAP varies far more widely during dialysis than previously recognized. This variation is associated with large changes in Qa that may impair accuracy in predicting thrombosis. This wide MAP variation also indicates hemodynamic reproducibility is not feasible when measuring Qa. Thus, we do not recommend standardized timing of Qa measurements during dialysis. A practical method of addressing poor Qa reproducibility may be to take frequent measurements so that trends can be recognized before thrombosis occurs.

Blood Flow Surveillance of Hemodialysis Grafts and the Dysfunction Hypothesis

It is widely recommended that hemodialysis graft surveillance programs should be implemented and that significant stenosis should be corrected when it is accompanied by graft dysfunction. The rationale for surveillance depends on the dysfunction hypothesis, which states that stenosis causes graft dysfunction [such as a decrease in graft blood flow (Qa)], and this dysfunction reliably precedes and accurately predicts thrombosis. The usefulness of Qa surveillance depends on accurate prediction of thrombosis so that stenosis can be corrected prior to thrombosis. An analysis of the dysfunction hypothesis indicates that some or all of its underlying assumptions are invalid. Most importantly, the presence of wide hemodynamic variation during Qa measurements makes Qa a relatively inaccurate predictor of thrombosis. A number of studies have evaluated the value of surveillance with intervention in reducing thrombosis rates and prolonging graft life. Review of these studies show that few have been prospective and randomized, and many have included historical control groups. It is debatable whether these studies have established that Qa surveillance with intervention should be applied to all grafts. Data from several studies suggest that severity of stenosis may be at least as accurate as Qa in predicting thrombosis. Consequently, inclusion of stenosis measurements (e.g., by duplex ultrasound) may improve the results of surveillance. These unresolved issues indicate it is premature to recommend routine Qa surveillance with intervention of all hemodialysis patients with grafts.

Hemodynamics of the Hemodialysis Access: Implications for Clinical Management

Optimum prevention and management of access complications require an understanding of access hemodynamics. The hemodialysis access is unique in that it creates a low resistance shunt between the arterial and venous circulations. Thus, it is unlike the arterial synthetic graft used in atherosclerotic vascular disease in which the arterioles continue to provide the main source of resistance and blood flow regulation. This chapter describes access hemodynamics for nephrologists, interventionists, surgeons, and others who are involved in the management and preservation of the hemodialysis access. It develops a hemodynamic model that provides insight into access complications and management. Definitions and units of measurement are listed in the Appendix.

In vivo validation of glucose pump test for measurement of hemodialysis access flow

BACKGROUNDnThe glucose pump test (GPT) is a recently introduced method of measuring hemodialysis access blood flow (Qa). A validation of GPT during dialysis has not yet been done, and performance characteristics of the method have not yet been fully analyzed.nnnMETHODSnThe authors studied 33 patients (25 synthetic grafts, 8 autogenous arteriovenous fistulae). Qa measurements by ultrasound dilution (UD) and GPT were done in triplicate during dialysis. In GPT, a baseline blood sample (C(1)) was obtained, followed by infusion of a 10% glucose solution (C(i)) through the arterial needle into the access at 16 mL/min (Q(i)). After 11 seconds, a downstream blood sample (C(2)) was aspirated from the venous needle. C(1) and C(2) glucose were measured by glucometer. Qa was computed by the equation: Qa = Q(i)(C(i) - C(2))/(C(2) - C(1)). A model of the access vascular circuit was used to determine the influence of C(2) aspiration on the Qa measurement.nnnRESULTSnMean Qa was 1413 mL/min by UD versus 1,496 mL/min by GPT (P = 0.11). There was a strong linear correlation between the 2 methods (r = 0.905; P <0.001). The pooled coefficient of variation was 6.4% for UD and 9.6% for GPT. The circuit model showed that aspiration of C(2) causes an increase in Qa (DeltaQa) that depends on the aspiration rate (Q(ASP)) and fraction of resistance in the circuit that is downstream to the venous needle: DeltaQa = Q(ASP)(Downstream resistance)/(Total resistance). The model predicts the overestimate is approximately 62 mL/min for grafts and 120 mL/min for fistulae but may vary depending on the balance of resistances upstream and downstream to the venous needle.nnnCONCLUSIONnThis study shows that GPT closely correlates with UD, and the method has adequate precision. GPT is an inexpensive method that may help make Qa measurements more widely available than previously possible.

Blood flow surveillance of hemodialysis grafts: insights from two case reports.

It is widely recommended that all hemodialysis grafts undergo blood flow (Qa) surveillance, and that stenosis be corrected when accompanied by a low Qa or decrease in Qa (ΔQa). This recommendation has, however, become increasingly controversial. Studies have shown that although there is an association between Qa and thrombosis, the accuracy of Qa in predicting thrombosis within individual patients is poor. We describe two cases that demonstrate common causes of poor predictive accuracy. These cases also show that application of Qa surveillance algorithms is often complex and ambiguous. Most studies reporting that surveillance with intervention reduces thrombosis or prolongs graft life have used historical or sequential control groups, or have been retrospective. Accurate assessment of the benefit of graft surveillance must await studies that are fully prospective and randomized with concurrent control groups. Until such studies have demonstrated sufficient benefit, we do not recommend periodic Qa surveillance with intervention of all hemodialysis grafts.

Fellows' Forum in Dialysis edited by Mark A.Perazella: Blood Flow Surveillance of Hemodialysis Grafts: Insights from Two Case Reports

It is widely recommended that all hemodialysis grafts undergo blood flow (Qa) surveillance, and that stenosis be corrected when accompanied by a low Qa or decrease in Qa (ΔQa). This recommendation has, however, become increasingly controversial. Studies have shown that although there is an association between Qa and thrombosis, the accuracy of Qa in predicting thrombosis within individual patients is poor. We describe two cases that demonstrate common causes of poor predictive accuracy. These cases also show that application of Qa surveillance algorithms is often complex and ambiguous. Most studies reporting that surveillance with intervention reduces thrombosis or prolongs graft life have used historical or sequential control groups, or have been retrospective. Accurate assessment of the benefit of graft surveillance must await studies that are fully prospective and randomized with concurrent control groups. Until such studies have demonstrated sufficient benefit, we do not recommend periodic Qa surveillance with intervention of all hemodialysis grafts.

Applicability of an entry flow model of the brachial artery for flow models of the hemodialysis fistula

The native arteriovenous fistula creates a shunt that provides the high blood flow that is needed for dialysis. Lumped parameter hemodynamic models of the arteriovenous fistula can be used to predict shear stresses and pressure losses and can be applied to help understand unsolved problems such as the high rate of arteriovenous fistula maturation failure. These models combine together flow components, such as arteries, stenosis, anastomoses, arterial compliance, and blood inertia, and each component must be modeled with an appropriate pressure–flow relationship. Poiseuille flow is generally assumed for straight vessels, but the unique high flow rates within the brachial artery of an arteriovenous fistula are expected to induce entry flow effects that are neglected in this model. To estimate the importance of these effects, brachial artery flow was modeled in a low-resistance network, such as the one that occurs when an arteriovenous fistula is constructed, through the lumped parameter model, and the predicted flow rates and pressures were compared to those predicted by computational fluid dynamics. When Poiseuille flow was assumed, the flow rate from the lumped parameter model was consistently larger than that from computational fluid dynamics, with a cycle-averaged error of 36.8%. When an entry flow model (Shah) was assumed, the lumped parameter–based flow was 6% lower than the computational fluid dynamics model at the peak of the flow waveform, and the cycle-averaged error was reduced to 7.8%. Thus, in a low-resistance (high flow) arteriovenous fistula circuit, an entry flow model can account for steeper near-wall velocity gradients. This result can provide a useful guide for designing engineering models of the arteriovenous fistula.