The effect of two different protocols of potassium haemodiafiltration on QT dispersion

Background. The risk of developing cardiovascular diseases is higher in patients on haemodialysis than in the general population. These patients may develop arrhythmias that depend on the extra- and intracellular concentrations of potassium. ECG findings, particularly the QT interval and its dispersion (QT(d)) and the QT(c) (QT interval corrected for heart rate according to Bazett's formula) and its dispersion (QT(cd)), may be direct indicators of the risk of developing arrhythmia. Methods. Our cohort comprised 28 patients who were dialysed for 3.5-4 h three times per week, first with haemodiafiltration with a constant potassium concentration (HDF) in the dialysis bath then with haemodiatiltration with variable concentrations of potassium (HDFk). ECGs were done at different time intervals: at the start of dialysis (T-0), at 15 (T-15), 45 (T-45), 90 (T-90) and 120 min (T-120) after the beginning of the session, and at the end of treatment (Tend). ECG-derived data (QT, QT(d), QT(c) and QT(cd)) were measured. At the same time points, plasma electrolytes, intra-erythrocytic potassium and the electrical membrane potential at rest (REMP) of the erythrocytic membrane were measured. Results. Plasma potassium concentration diminished more gradually in HDFk than in HDF, the difference being statistically significant at T-15 and T-45 (P<0.05), and T-90 (P<0.01). The intra-erythrocytic potassium concentration remained constant throughout the observation period. In both HDF and HDFk, REMP was lower at all points after T-0 (P < 0.05), but the reduction was greater and more significant in HDF than in HDFk at T-15 and T-120 (P<0.05). ECG revealed a statistically significant diminution in HDFk vs HDF in the measures of dispersion of QT and QT(c) at T-15, T-90, T-120 and T-end (P<0.01) and of QT(cd) at T-45 (P<0.05). The mean of QT(d), adjusted for plasma potassium, increased over time in HDF with large alternate mean increase and decrease peaks and error intervals. In HDFk, instead, there was a progressive and constant diminution with minor error intervals. QT(cd) adjusted for plasma potassium had the same trend. A marked difference was found between the final values in standard HDF and those in HDFk. Conclusions. HDF and HDFk have significantly different effects on QT(c). ECG data demonstrate that the risk of arrhythmia could be lower, with a variable removal of potassium during haemodialysis. With HDF but not HDFk, hyperpolarization of the cell membrane is detected, and this could have a destabilizing effect on different types of cardiac cell, giving rise to retrograde circuits.