Received 18 December 2015; accepted 26 February 2016; published 29 February 2016
Although these days a wide range of progressions take place in dialysis technology, cardiovascular diseases remain as the single most common cause of death in chronic dialysis patients  . One of the major causes of death in end-stage renal disease (ESRD) patients under maintenance haemodialysis (HD) is ventricular arrhythmias  .
Although uremic patients usually already have a decreased potassium pool   , in order to counterbalance the interdialysis potassium load and avoid life-threatening hyperkalaemia, potassium is removed by HD. Unlike HD sodium removal, which can even be exclusively convective   , HD potassium removal is almost exclusively diffusive. It is therefore necessary to create a gradient between plasma and dialysate potassium concentrations, which means greatly reducing plasma potassium concentrations during HD. However, very different amounts of potassium removal can be obtained with quite small differences in the dialysate plasma gradient since the potassium removal is mainly due to a decrease in the intracellular potassium pool.
However, relatively few studies have examined patient and HD-specific factors that might be associated with a higher risk of developing cardiac arrhythmias. The arrhythmogenic effect of HD was considered not only because of its still discussed clinical importance, but also because it is a “marker” of the electrophysiological status of the cells of uremic patients, which are greatly modified during standard HD treatment three times a week for many years  .
Hence, we have undertaken this study of arrhythmias in order to evaluate the role of one of the major factors in the genesis of HD arrhythmias, potassium (K) changes during dialysis.
We have evaluated the effects of two different regimens of dialysis potassium removal in patients with a tendency to develop arrhythmias during HD. Our chief interest was to identify the regimen of the least susceptibility to cardiac rhythm disorders during the dialysis cycle.
End-stage renal failure patients
There were 88 (36 men and 52 women) ESRD patients recruited for the study. They received regular haemodialysis three times per week at the haemodialysis units of a university medical centre (Golestan hospital) during year 2011. All the patients have received a written informed consent approved by Ahwaz Jondishapour research ethic board, and they all agreed to participant the research. Clinical indication for haemodialysis in the present study population is divided into two categories: 1) absolute indication with creatinine clearance rate (Ccr) < 5 ml/min or serum creatinine (Cr) > 8.0 mg/dl; 2) relative indication with Ccr < 15 ml/min or serum Cr > 6 mg/dl and with accompanying life threatening complications as congestive heart failure, lung edema, haemorrhage diathesis, consciousness change, cachexia or uncontrollable hyperkalaemia with drugs. Uremic patients on stable HD treatment were recruited. After they had given their oral consent, the patients fulfilling the entry criteria were selected for the experimental phase.
Patients aged more than 18 years, who had been on chronic thrice-weekly HD for at least six months, and who were in stable clinical condition were admitted to the study. The continuous 24 hours electrocardiographic (ECG) recording made on the HD day was performed for each patient.
Patients on antiarrhythmic treatment or receiving pace-maker cardiac stimulation, those whose dosage of digitalis was unstable, and those requiring dialysate K concentration of more than 3 mEq/litter were excluded from the study.
The patients had to undergo HD at the same time of day for both treatments in order to avoid any interference with the circadian rhythms of arrhythmia. In patient and throughout the three weeks of the study, only the concentration of dialysate K could change, and then only in accordance with the study protocol.
Twenty-four hour EGG recording on the HD day, continuous 24-hour EGG recording was started at the time of HD. The 24-hour EGG signal was recorded and the tapes were centrally analysed for ventricular arrhythmias.
The readers were blinded as to treatment.
All uremic patients that had inclusion criteria randomly divided into two groups. Group A (consist of 44 patients) Standard HD with a dialysate K concentration of 3 mEq/litter. Group B (consist of 44 patients) Experimental HD with dialysate K2 mEq/lit concentration were performed for them.
The calculated sample size of 88 patients was based on: a significance level (an error) of 0.05; a β error of 0.2; a power [1-β] of 0.8; after gathering data they were analyse by using SPSS soft ware version 16 and ANOVA and student T test was used for analysing data.
A Holter recording selection some of the 24-hour Holter recordings were considered totally unsuitable for analysis because of the bad quality of the ECG signal (presence of superimposed noise, incorrect positioning of the electrodes) and the consequent inability to distinguish normal beats from pathological PVCs. The bad signal quality was sometimes present during only a part of some of the recordings, leading to temporary gaps in which the PVCs could not be detected.
88 ESRD patients (36 male) were recruited in the study. The mean age of patients was 58.6 ± 8.3 (Table 1).
As its shown in Table 2 the mean age of patients and the electrolyte levels of two groups of ESRD patients [with dialysate potassium solution of 2 mEq/lit (k2) and Dialysate potassium of 3 mEq/lit (k3)] were comparable and had not significant change (P > 0.05).
The serum electrolyte levels of NA, K, BUN, CR, CA, P, PH after dialysis were decreased significantly (P < 0.05) (Table 3).
As shown in Table 4 it was a trend for PVC, AF to increase with potassium solution of 2 mEq/lit (numbers of PVC, AF in patients with potassium solution of 2 mEq/lit was 37.7 ± 6.1, 13.28 ± 0.5 and in patients with potassium solution of 3 mEq/lit these were 12.4 ± 1.6, 9.2 ± 0.4). Significant change was seen with K2 in the number of SVC.
Table 5 shows an increase in the difference of potassium before and after dialysis lead to significant change in the number of PVC and SVC.
Table 1. Age and sex distribution of end stage renal disease patients.
Table 2. Electrolyte and age distribution in different dialysis potassium solution.
Table 3. Electrolytes changes before and after hem dialysis in ESRD patients.
Table 4. Numbers and standard deviation and confidence interval of PVC and AF and SVC in two different potassium solutions.
Table 5. Mean potassium differences in plasma of ESRD patients in different dialysate potassium solution and number of PVC and SVC in them.
In HD with dialysate potassium solution of 3 mEq/lit, a lower appearance of PVC as compared to dialysate potassium solution of 2 mEq/lit was apparent, even though the difference did not turn out to be statistically significant (P = 0.09), but SVC appeared more in group with dialysate potassium solution of 2 mEq/lit no significant difference was found between the two treatments in either isolated PVC, or pairs or runs. Again, there was no statistical difference between the two treatments in regarding to electrolytes.
Studies on dialytic potassium removal   show that the higher the plasma dialysate potassium gradient, the greater the potassium removal from intracellular fluid. The uraemia related inhibition of the Na-K pump produces a potassium shift from intra-to extracellular fluid and consequently induces hyperkalaemia; thus, the more advanced the cellular sickness, the bigger the cellular depauperation of Potassium during analysis. Consequently, there is a vicious circle: although it improves the function of the Na-K pump, dialysis treatment maintains and even increases cellular distress by reducing the intracellular potassium pool. Thus, in our opinion, the new approach explored in this study seems to be particularly suitable for patients prone to life-threatening hyperkalaemia  - .
It is known from physiology studies that the wholeness of the cellular potassium pool is necessary to many cellular functions: volume and PH regulation  .
Given selective cell permeability to potassium (K), the resting electric membrane potential (REMP) of most cells is related to the diffusive passive fluxes of K ions  . The level of REMP affects cell excitability: the greater the difference between REMP and the threshold level of the electric potential, the longer the depolarization time   .
In chronic haemodialysis (HD) patients, intradialytic K is removed by diffusion throughout the HD membrane, according to the concentration gradient between plasma and dialysate K levels. With constant dialysate K levels of 2 mEq/lit, this gradient decreases rapidly during the first hour of HD and slowly in the following hours. The diffusive fluxes of K through the HD membrane produce diffusive passive fluxes through the membrane cells, with greater negativization of membrane of cells and, consequently, less cell membrane excitability. This phenomenon is more evident during the first hour of HD, since the plasma-dialysate K gradient is higher.
The changes in K fluxes induced by HD may therefore influence cardiac cell electrophysiology, and it may lead to cardiac arrhythmias.
In HD with constant and low potassium (range 2 mEq/l) a large amount of potassium is abruptly removed from the extracellular space  .
The depletion of the potassium reserves within the cells may have important repercussions on cardiac electrophysiology.
Another mechanism by which potassium change lead to cardiac arrhythmia is that Potassium fluxes during HD have been associated with an increase in QT interval   , an increase in the dispersion of QT and in the inhomogeneous repolarisation revealed by the analysis of the spatial aspects of T-wave complexity  . The resulting repolarisation heterogeneity allows for the onset of distinctive re-entrant arrhythmias, and hypokalaemia may act as a triggering factor in the genesis of premature ventricular depolarisations. The incidence of related ECG abnormalities during HD ranges from 18% to 76%, depending on the definition of the abnormal ventricular electrical activity  .
An Italian study by Redaelli et al.  has demonstrated a 36% reduction in premature ventricular complexes using a profiling of potassium dialysate concentration in arrhythmic HD patients. Morrison et al.  demonstrated decreased ventricular ectopic activity in four out of six patients whose dialysate potassium concentration had changed from 2 to 3.5 mEq/lit. In our study, the serum potassium trends are significantly different with the use of constant K concentration (3 mEq/l) in comparison with (2 mEq/l) K dialysate concentration and in our study the number of PVC increased more than 3 times in dialysate solution with potassium 2 mEq/l that was very close to the result of Redaelli et al. on the other hand SVC numbers raised more than 5 times independently.
Rombol`a et al.  found like us a greater fall in intra-erythrocyte potassium concentration in patients with an arrhythmic tendency, as compared to those without such a tendency  but he didn’t exactly mention witch kind of arrhythmia are more prone to the potassium changes.
In our opinion analysis of data support Rombol`a idea that some stressors such as fluid overload, increase in blood pressure and the changes in pH and bicarbonates, may act as trigger of ventricular arrhythmias  . Obviously this may be particularly true for patients with cardiomyopathy, left-ventricular hypertrophy, coronary artery disease and cardiac heart failure, which provide the perfect backdrop for arrhythmias to occur  .
Naturally, these demonstrate a higher incidence of sudden death in a group of patients treated by HD with constant 2 mEq/l potassium dialysate in comparison with another group treated with constant 3 mEq/l potassium solutions.
In conclusion, the use of a model of intra-HD potassium that is more close to potassium serum concentration of ESRD patients can decrease the arrhythmogenic effect of HD in patients on regular HD treatment. This result was obtained without adversely affecting pre-dialysis plasma potassium levels in comparison with the standard HD procedure. The results of this study not only showed that the amount of cardiac arrhythmia raised more in potassium concentration of 2 mEq/lit but also demonstrated that supra ventricular arrhythmias are more influenced by those changes. The results of this study clearly illustrate that the use of a model of HD potassium removal which is more close to blood concentration of potassium is capable of reducing the arrhythmogenic effect of standard HD in chronic uremic patients prone to this complication. This decrease was statistically and clinically significant for both PVC and SVC.