Blood volume regulation during hemodialysis

Hemodialysis (HD)-induced hypotension may be precipitated by severe hypovolemia. To avoid the appearance of destabilizing hypovolemias, we have developed a biofeedback control system for intradialytic blood volume (BV)-changes modeling. The system, incorporated in a dialysis machine, is based on a multivariable closed-loop control with a dependent output variable, the BV changes, and two independent control variables, the ultrafiltration rate (Qf) and dialysate conductivity (DC). The relative BV changes occurring during HD are measured by an optical device. The Qf and DC are continuously adjusted by the control model during the treatment to minimize any discrepancies between the ideal targets for the BV, the patient's body weight reductions, and the experimentally obtained results. The system manages three kinds of errors: in BV changes, the total weight loss, and the sodium balance. The latter is controlled by a dedicated kinetic model that continuously calculates the equivalent DC and, by the end of the session, tends to make the sodium balance the same as the one obtained in conventional HD with constant DC. This system's capacity to improve intradialytic hemodynamic tolerance has been assessed in a crossover study of eight highly symptomatic patients. Conventional HD (CHD; period A) was compared with blood volume-controlled dialysis sessions (BV-CHD; period B) following a protocol with an A1-B-A2 sequence, with each period lasting 1 month. A lower decrease in BV (-10.6%) was obtained during BV-CHD (period B) compared with CHD (-12.3% in period A1 and -12.5% in period A2). The predialysis to postdialysis systolic arterial pressure changes were lower in period B (-12.4%) than in period A (-20% in A1 and -17.5% in A2; P < 0.05) despite similar total Qf and mean treatment times. A significant reduction in the number of severe hypotensive episodes (three in period B v 26 in period A1 and 16 in period A2; P < 0.05) and the overall incidence of complaints, especially of muscular cramps, was found in BV-CHD. These results were reflected in a reduced need for therapeutically administered isotonic saline in each session (60 mL in B v160 mL in A1 and 95 mL in A2; P < 0.05). In conclusion, the proposed biofeedback system for intradialytic BV control may be useful to avoid severe hypovolemic states, to stabilize BV by modeling its trend, and to avoid reaching individual critical BV thresholds in hypotension-prone patients.