Hypokalemia is found in 1% of healthy people who do not take any medication. Even though during one hemodialysis session patients lose larger amounts of potassium (K⁺) compared to Peritoneal Dialysis (PD) patients [1], hypokalemia in PD patients is more common (10% - 36%) [1]-[6]. Another important difference between hemodialyzed and PD patients is the fact that in the hemodialyzed K⁺ is removed quickly, causing large changes in the blood levels, while in PD patients the dialytic removal of potassium is slower [6].

Hypokalemia in PD patients is attributed to many causes. First, in the dietary intake, both due to the content of food (in those patients it is often poor in K⁺), and to the way of cooking [7]. Of course, it may be owing to the redistribution, where the continuous secretion of insulin contributes, due to the presence of glucose [8], but also due to alkalosis [9], which moves K⁺ intracellularly. It may also be due to skin and renal losses (in those with residual renal function) [7], but also to peritoneal losses, in those who are treated with more than 4 PD exchange bags/24 hours (where there are continuous potassium losses by diffusion) [7], [10]. Some medications may also affect blood K⁺ levels, some of which insert K⁺ intracellularly (i.e., insulin, catecholamines, thyroxine) and others that prevent the transcellular movement of K⁺ (i.e., β-blockers, heparin), or that do not allow its renal or intestinal elimination from the body (angiotensin-converting enzyme inhibitors-ACE, angiotensin receptor blockers-ARB, etc.). However, hypokalemia is generally found in PD patients who are treated with higher volumes of PD fluid/24 hours (i.e., those who are exposed to higher concentrations of glucose, but also in those who achieve higher doses of clearance) [6][11][12].

In this prospective study, we recorded patients with moderate to severe hypokalemia who were treated with solutions given intraperitoneally with significantly greater concentrations of K⁺ than those reported in the literature.

Among 62 patients who were under PD during a period of 2 years, 10 with moderate to severe hypokalemia (K⁺ from 2.1 to 3.0 mmol/L) were recorded. Patients’ age, sex, type of peritoneal dialysis (automated or continuous), number of PD exchange bags/24 hours, ultrafiltration volume/24 hours, type of peritoneal membrane transport characteristics, kind of bag solution, existing residual renal function, presence or absence of diabetes, and existence of malnourishment are shown in Table 1.

All patients were taking β-blockers, heparin 2/10, NSAIDs 3/10, eplerenone 3/10, and ACE inhibitors 3/10. All had severe hypoalbuminemia (serum albumin 29 ± 7 g/L, range from 13 - 35 g/L). In terms of symptoms, only 2/10 had weakness and discharge, one had anorexia, and one had a headache. No one had any acid-base disorder (pH normal range from 7.39 - 7.49), and their blood bicarbonate levels ranged from 20.9 - 32.3 mmol/L.

In our patients with hypokalemia, 3 - 6 exchanges of 2 L PD bags with a content of K⁺ 20 mmol/L had to be made to restore the K⁺ of the blood to levels higher than 4 mmol/L. Only 2 patients complained of pain when the infused fluid was in the abdomen (in the second one, after well stirring of the bag, no pain reappeared). No case of hyperkalemia was noted after the infusion of the exchange bag. The number of PD exchange bags needed for each patient to restore the blood levels of K⁺, the blood K⁺ levels, and the glucose concentration after each change are shown in Table 2.

Table 2.Changes in blood K⁺ levels and glucose levels after each exchange bag of PD solution.

Serum K⁺ levels are relatively stable in PD patients, compared with hemodialyzed patients [1]. However, many studies have shown that hypokalemia is more common in patients undergoing PD (PD solutions do not contain K⁺) than in those on hemodialysis [3][8][13][14]. Pai et al. recorded symptomatic hypokalemia in 5% of their PD patients who had to visit the emergency department (ED) [15], while Szeto et al. found serum K⁺ < 3.5 mmol/L in 20% of their patients [16]. In general, however, 10% - 30% of patients on PD need K⁺ supplementation [2][16]-[18].

Dietary potassium contributes to its balance in the body. Of course, malnutrition is also associated with hypokalemia [16], as we have seen from the elevated serum C-Reactive Protein (CRP) levels in hypokalemic patients and in those with decreased serum albumin levels [5][13][16][19][20]. Others have found a link between hypokalemia and malnutrition [5][16], including us, as our patients (7/10) were clinically malnourished and all had serum albumin < 3.5 g/L (range from 1.9 to 3.5 g/L). It is noted that malnutrition, which is relatively common in PD patients, when maintained for months, is also associated with an increased incidence of peritonitis, because it is accompanied by overhydration, which leads to swelling of the intestinal mucosa, which is associated with easier endotoxin passage through the intestinal wall [13][21]. This may be true because hypokalemia is associated with decreased intestinal motility [16][20]. In the diet of malnourished patients, it should be borne in mind that each gram of protein contains 1 mmol of K⁺ [5][17] and that malnutrition indicates reduced food intake [5].

Additional factors that may potentially have an influence on serum potassium levels in PD patients include the influence of glucose-containing PD solutions on transcellular movement of potassium. Hypokalemia may therefore be due to redistribution, because of the continuous insulin secretion due to the presence of glucose [8], which is confirmed by muscle biopsies that showed a higher concentration of K⁺ in PD patients compared to those on hemodialysis [22], something that others disagree with [4]. Of course, no correlation was found between the incidence of hypokalemia and the presence of diabetes mellitus [5], something that we also found, since 7/10 of our patients did not have diabetes. However, what should be emphasized is that even though many of our patients were taking medications that do not allow intracellular movement of K⁺ because they do not allow Na⁺-K⁺-ATPase to function properly (10/10 patients received β-blockers and heparin 2/10), they had low blood K⁺ levels, which means that the hypokalemia in some of them was more significant.

Also, catecholamines and thyroxine, which stimulate Na⁺-K⁺-ATPase, may be responsible for hypokalemia; however, there were no such cases among our patients. Redistribution can, of course, be caused by blood acidification with a negative effect on Na⁺-K⁺-ATPase (for each decrease in pH by 0.1, the serum K⁺ increases by 0.8 mmol/L) [9].

Potassium is removed by PD with diffusion [13][23]; therefore, removal is affected by blood K⁺ concentration [18]. PD removes 1.5 K⁺ mmol/hour [24], and 4 PD exchanges bags of 2 L remove 20 - 30 mmol K⁺ daily [13][25], an amount that is quite less than that taken in by food from each patient/24 hours (with the Western diet, people take in on average 60 - 90 mmol K⁺/24 hours) [12].

Another cause of hypokalemia in PD patients is an increase in the clearance provided [7][10] and an increase in ultrafiltration with solutions with high glucose concentration [6][11][12]. Most of our patients (8/10) were treated with 4 PD exchange bags of 2 L/24 hours, one with 5 exchange bags of 2 L, and one was in APD with 11 L dialysate/24 hours, so there is no question of increased clearance. The glucose concentration of the PD bags affects insulin secretion and the redistribution of K⁺ between intracellular and extracellular spaces [14], so it also affects blood K⁺ levels (perhaps in 2 of our patients this parameter played a role, since one was on 4 and the other on 1 semi-hypertonic PD exchange bag).

Icodextrin dialysate can affect serum K⁺ because it increases PD ultrafiltration (which means an increase in serum K⁺ concentration) [13], does not cause hyperglycemia like glucose solution, and improves intestinal motility, which contributes to better dietary food intake and K⁺ [26][27] and does not affect appetite negatively, as conventional glucose dialysate solutions do. Yi et al. studied serum K⁺ levels for 2 years in 255 patients on PD (116 on overnight glucose solution and 139 on icodextrin dialysate). The incidence of hypokalemia noted in the two groups was similar, with mean serum K⁺ levels being higher with icodextrin dialysate compared with glucose solution (p < 0.05). This study showed that icodextrin in chronic use during the night exchange is associated with increased serum K⁺ levels and in these individuals, hypokalemia is less common compared to those who use only glucose solutions. This effect of icodextrin dialysate is attributed to better nutritional status than to redistribution of K⁺ into the intracellular space [14]. We do not agree with these results, since 8/10 of our patients received icodextrin and nevertheless had a significant degree of hypokalemia.

Hypokalemia in our patients may also have been due to urinary loss of K⁺ (in patients with residual renal function) [7]. This may have been present in some of our patients, as 4/10 had residual renal function.

Some studies have found that ACE inhibitors affect serum K⁺ levels in PD patients [2][28][29], as well as aldosterone receptor inhibitors (3/10 were taking eplerenone). This is thought to occur through regulation of intestinal K⁺ excretion [30]. Specifically for spironolactone, it was found to be associated with a significant increase in serum K⁺ after 2 months of administration, without any side effects [31]. At the same time, Valencia et al. used spironolactone in PD patients who became hypokalemic because of diarrhea at a dose of >65 mg/24 hours and found a change in serum K⁺ after 12 months of treatment, with 20% developing hyperkalemia [32], although others did not find any change in serum K⁺ in 12 PD patients with hypokalemia who received 25 mg spironolactone/24 hours [33]. The effectiveness of spironolactone in improving hypokalemia is attributed to its effect on the epithelial cells of the colon, where it inhibits the receptors of mineralocorticoids, preventing the loss of K⁺ through the large intestinal epithelium [34]. Of course, when there is residual renal function, the action of spironolactone in the distal renal tubules is added, with an additional benefit in improving K⁺ levels.

Despite the small number of patients in our study, it is impressive that everyone was taking β-blocker, 3/10 eplerenone, 2/10 heparin, 3/10 NSAIDs, 3/10 ACE inhibitors, and 8 icodextrin dialysate, each of which could somehow increase blood K⁺ concentration, but all of them had moderate to severe hypokalemia.

The low blood K⁺ concentration can affect the myocardial membrane resting potential, repolarization, and stimulus conduction velocity, which has an impact on mortality. Severe hypokalemia in PD patients may be manifested by muscle disorders (muscle weakness, rhabdomyolysis) [17], cardiac arrhythmia, weak peripheral pulse, orthostatic hypotension, decreased cardiac output [35], and in the intestine with constipation, which favors the occurrence of peritonitis episodes [35][36].

In PD patients, low serum K⁺ levels are positively correlated with all causes of death, but also with cardiovascular events [6]. It is also associated with sudden deaths in patients with cardiovascular disease [37]. In fact, hypokalemia has been found to be an independent prognostic risk factor for survival [16] and a risk factor for cardiovascular causes of death [38]. Hypokalemia in PD patients is generally associated with reduced survival [5][16]. In fact, a large study of 10,468 PD patients revealed that low serum K⁺ levels contributed significantly to the occurrence of death in them, compared to hemodialyzed patients. It was also found that patients with serum K⁺ < 4 mmol/L were 2 times more likely to die than those with serum K⁺ > 5.5 mmol/L [38].

Xu et al. considered that patients on PD with serum K⁺ 3.0 - 3.5 mmol/L can restore their levels with a diet rich in K⁺, while when the blood levels are 2.5 mmol/L intravenous administration is required [6]. This is important because hypokalemia is dangerous for the development of arrhythmias and is associated with sudden deaths, while restoring serum potassium with foods rich in K⁺ is not as easy and as fast as required.

In Australia, it is recommended to administer a PD exchange bag with a K⁺ concentration of 3 mmol/L when the serum K⁺ is 3 - 5 mmol/L and a 4 mmol/L solution when the serum K⁺ is <3 mmol/L. In our country, in general, the dose of K⁺ given intraperitoneally to PD patients with hypokalemia corresponds (usually 5 ml of 10% KCI is added to every 2 L of PD solution). In contrast to this conservative treatment, Spital & Sterns, who gave intraperitoneal K⁺ to five PD patients with hypokalemia in only one PD exchange bag of 2 L, found that it was well tolerated at a density of 20 mmol/L, while administration of a solution of 40 mmol/L in one PD exchange bag caused severe abdominal pain in some of them [2]. We administered a PD solution with 20 mmol/L K⁺, continuously in each exchange bag until the K⁺ of the blood concentration became > 4 mmol/L, which restored the blood K⁺ levels gradually, without significant side effects (abdominal pain during infusion of the dialysate was found in only two of our patients, and after good mixing of the solution in one of them it did not recur), with no hyperkalemia being found in any of our patients.

Amirmokri et al. were, of course, more daring and, in 5 patients on Continuous Ambulatory Peritoneal Dialysis (CAPD) with severe hypokalemia, they gave K⁺ intraperitoneally at a dose of 60 - 80 mmol K⁺ in a 2 L PD exchange bag, i.e., 30 - 40 mmol/L. Specifically, K⁺ was added to a PD dialysate of 2.5% glucose during the night exchange, and serum K⁺ was determined after 2, 4, 6 - 8, and 20 - 24 hours of solution infusion in the abdomen. They found an increase in serum K⁺, on average from 3.2 mmol/L (range 2.6 - 3.6 mmol/L) to 4.1, 4.2, 4.1, and 4.2 mmol/L, respectively, during the reported blood sampling hours [23], without side effects. It is noted that such a high concentration of K⁺ in a PD bag was administered for the first and last time.

It is concluded that intraperitoneal administration of K⁺ at a concentration of 20 mmol/L in each PD exchange bag, until the concentration of K⁺ increases to levels greater than 4 mmol/L, is safe and is not accompanied by any side effects.