Model-based analysis of potassium removal during hemodialysis.

Potassium ion (K(+)) kinetics in intra- and extracellular compartments during dialysis was studied by means of a double-pool computer model, which included potassium-dependent active transport (Na-K-ATPase pump) in 38 patients undergoing chronic hemodialysis. Each patient was treated for 2 weeks with a constant K(+) dialysate concentration (K(+)(CONST) therapy) and afterward for 2 weeks with a time-varying (profiled) K(+) dialysate concentration (K(+)(PROF) therapy). The two therapies induced different levels of K(+) plasma concentration (K(+)(CONST): 3.71 +/- 0.88 mmol/L vs. K(+)(PROF): 3.97 +/- 0.64 mmol/L, time-averaged values, P < 0.01). The computer model was tuned to accurately fit plasmatic K(+) measured in the course and 1 h after K(+)(CONST) and K(+)(PROF) therapies and was then used to simulate the kinetics of intra- and extracellular K(+). Model-based analysis showed that almost all the K(+) removal in the first 90 min of dialysis was derived from the extracellular compartment. The different K(+) time course in the dialysate and the consequently different Na-K pump activity resulted in a different sharing of removed potassium mass at the end of dialysis: 56% +/- 17% from the extracellular compartment in K(+)(PROF) versus 41% +/- 14% in K(+)(CONST). At the end of both therapies, the K(+) distribution was largely unbalanced, and, in the next 3 h, K(+) continued to flow in the extracellular space (about 24 mmol). After rebalancing, about 80% of the K(+) mass that was removed derived from the intracellular compartment. In conclusion, the Na-K pump plays a major role in K(+) apportionment between extracellular and intracellular compartments, and potassium dialysate concentration strongly influences pump activity.

A method for monitoring vascular access function during hemodialysis.

We tested a new bedside method to determine the function of native arteriovenous fistula in 16 patients performed during hemodialysis without stopping the treatment. We initially measured vascular access flow (Q(a)) in each patient using the Transonic HD01(plus) device. We then measured the pressure in arterial and venous drip chambers at different blood pump flow rates (Q(bset)=0, 50, 100, 250, 300, 350 ml/min). The intravascular blood pressure gradient (P(f)) between arterial and venous puncture sites was estimated by a mathematical model. P(f) was positive for low Q(bset), but became negative when Q(bset) overcame the threshold value (Q(Inv)). Such critical flow showed a high correlation with Q(a), even if it was systemically lower. Computer analysis of fluid dynamics showed that when the blood pump flow overcame the Q(Inv) threshold, a critical transition from laminar flow to vortex circulation took place downstream of the venous needle, causing a dangerous shearstress on the vessel wall. Our results show that Q(Inv) provides an indication of the maximal blood pump flow rate needed to be reached to maximize blood flow supply in order to limit hemodynamic stress on the vascular access.

Cardiac response to hemodialysis with different cardiovascular tolerance: heart rate variability and QT interval analysis.

A therapy-specific worsening of cardiovascular stability during bicarbonate dialysis (BD) with respect to acetate-free biofiltration (AFB) have been previously reported. We further investigated the impact of the 2 therapies on electrocardiographic parameters in order to gain novel insight into the cardiac responses. Holter ECG acquired during hypotension-free sessions (12 BD + 12 AFB) were retrospectively analyzed. R-R intervals were extracted from ECG recordings. An autoregressive spectral technique was used to compute low- and high-frequency (LF and HF) components of heart rate variability (HRV). QT interval duration was measured with a computer-assisted technique and corrected for HR. In BD the LF component of HRV after an initial increase was slowly depressed with respect to AFB (p < 0.05). QT duration showed a significant (p < 0.01) hemodialysis-induced reduction. QT shortening was more pronounced (p < 0.05) in BD than in AFB (-31 vs. -10 ms), even after correction for HR (p < 0.05). Cardiac electrical activity is significantly affected by the hemodialysis technique. The decrease in the LF component of HRV and the QT shortening are coherent with the worse cardiovascular tolerance observed in BD and with the hypothesis of an enhanced production of endogenous nitric oxide.

Mathematical modeling of arterial pressure response to hemodialysis-induced hypovolemia.

A computer model of pressure response to hemodialysis-induced hypovolemia is reported. Heart rate and hematocrit, measured in the course of hemodialysis, are imposed as computer model inputs and the model computes the arterial pressure response after tuning model parameters representative of patient's cardiovascular reactivity. Computer model reproduced with good accuracy experimental data (arterial pressure, cardiac output and total peripherical resistance). Parameter identification over successive sessions of the same patients revealed satisfactory reliability, providing a physiological interpretative key to patient's hemodynamic behavior during hemodialysis.

Model-based study of the effects of the hemodialysis technique on the compensatory response to hypovolemia.

BACKGROUND: Hemodialysis technique (dialysate composition, filter, convection/diffusion ratio, etc.) can have an impact on the patient's tendency to acute hypotension. We have examined the hypothesis that the dialysis technique affects the hypotension risk by altering the cardiovascular compensatory response to hemodialysis-induced hypovolemia. METHODS: Twelve hypotension-prone subjects were studied during six sessions of conventional bicarbonate dialysis (BD) and six sessions of acetate-free biofiltration (AFB). Blood volume (BV) control system was used in AFB to provide a BV change equivalent to the BV change observed in BD. The efficacy of reflex compensatory mechanisms was assessed by a model-based computer analysis of the BD and AFB sessions. RESULTS: BD sessions were complicated by hypotension more frequently than the AFB ones (34/66 BD vs. 18/66 AFB). Hypotension arose about 60 minutes earlier in BD (123 +/- 41 minutes in BD vs. 183 +/- 25 minutes in AFB, P < 0.01), and after a smaller BV reduction (hypotension BV 7.9%+/- 2.0% in BD vs. 10.9%+/- 2.6% in AFB, P < 0.05). Model-based computer analysis of the sessions without hypotension revealed differences in peripheral resistance adaptation (9%+/- 9% BD vs. 19%+/- 7% AFB, P < 0.05) as well as in the stroke volume reduction (19%+/- 8% BD vs. 10%+/- 8% AFB, P < 0.001). Model analysis of sessions with hypotension indicated that compensatory mechanisms were almost inoperative in BD, whereas a residual capacity to control peripheral resistance and cardiac contractility was present in AFB. Model simulations demonstrated that hypotension occurred later in AFB since the residual compensatory capacity in AFB was able to sustain the arterial pressure for larger BV reductions (8.3% BD vs. 11.2% AFB). CONCLUSION: The increased risk of acute hypotension in BD compared to AFB is caused by a therapy-induced inhibition of reflex compensatory response to hypovolemia.