#NephJC Chat
Tuesday November 19th, 2024 at 9 pm EST
Wednesday November 20th, 2024 at 9 pm IST
Kidney Int. 2024 Oct 26:S0085-2538(24)00722-1. doi:10.1016/j.kint.2024.10.010. Online ahead of print.
doi: 10.1016/j.kint.2024.10.010
Effects of dialysate potassium concentration of 3.0mEq/l with sodium zirconium cyclosilicate on dialysis-free days versus dialysate potassium concentration of 2.0mEq/l alone on rates of cardiac arrhythmias in hemodialysis patients with hyperkalemia
David M Charytan, Wolfgang C Winkelmayer, Christopher B Granger, John P Middleton, Charles A Herzog, Glenn M Chertow, James M Eudicone, Jeremy D Whitson, James A Tumlin ADAPT Investigators
PMID: 39490411
Introduction
The leading cause of death in dialysis patients is cardiovascular disease (CVD), typically from sudden death rather than ischemia. From the 2024 USRDS report, sudden death is the leading cause of mortality in dialysis, at about 47%. This may explain the suboptimal effect of statins in this population. For in-centre hemodialysis, the long interdialytic interval is associated with higher all-cause and cause specific mortality (Foley et al, NEJM 2011). In particular, the risk of dysrhythmias was almost twice as high after the long interdialytic interval (20.9 versus 11.1 per 100 person-years). What could be happening here? Apart from pulmonary edema, hyperkalemia is a probable contributor. One might say: ‘Let’s just do dialysis with a lower K bath and prevent hyperkalemia after the long interdialytic interval!’ Hold up, not so fast !
Causes of death, from 2024 USRDS report
The effects of adjusting the dialysate potassium and its ramifications on arrhythmias in dialysis patients has been previously evaluated. First, the use of a very low dialysate potassium, for example, 1.0 or even 0 mmol/L, has been associated with an increased incidence of sudden cardiac death (SCD) (Weisberg et al Semin Dial 2010). In another study (Lafrance et al, NDT. 2006), the authors examined the predictors and outcomes of cardiopulmonary resuscitation in prevalent hemodialysis patients. A low dialysate potassium bath was associated with an increased risk of cardiac arrhythmias and death. Similar findings were obtained from the DOPPS database (Jadoul et al CJASN. 2012) where it was noted that patients on hemodialysis being dialyzed with a 3 mmol/L potassium bath were least likely to develop a cardiac arrhythmia or SCD as compared to patients who were dialyzed with a potassium bath concentration of < 2 mmol/L. However, there might be confounding by indication since low potassium baths would more likely be used in those patients who had high pre-dialysis serum potassium.
So is it the high pre-dialysis serum potassium or the low dialysate potassium that is dangerous?
The answer might be found in a study of 502 hemodialysis patients compared with 1632 age/dialysis vintage matched controls. A low dialysate potassium (< 2 mmol/L), but not a high predialysis serum potassium, was associated with a sudden cardiac arrest in the dialysis unit. Interestingly, the low dialysate potassium was not protective in those with high predialysis serum potassium (Pun et al Kidney Int, 2010).
Image 1. Probability of sudden cardiac arrest (SCA) by dialysate potassium concentration (>2 meq/l, black line; <2 meq/l, red line) and last recorded serum potassium
From: Pun et al Kidney Int, 2010.
The purported explanation for this seeming paradox is that the fast shift of potassium (in those with a high preHD potassium on low K baths) and the consequent low post-dialysis potassium may be to blame for this increased risk. Naturally, we can’t leave peeps with high serum K to be dialyzed on a 3K bath. So, how do we reduce the serum predialysis K to allow a higher K bath? Potassium binding agents such as patiromer (Middleton et al. Kidney Int 2024) and sodium zirconium cyclosilicate (SZC) (Fishbane et al, JASN 2019) have recently been utilized for hyperkalemia in patients on dialysis, especially during the long interdialytic interval, and thereby permitting the use of higher dialysate potassium. Even the good old sodium polystyrene sulfonate could do this, but it seems to have fallen out of favor related to its poor palatability and potential GI adverse events. These K binders may help in mitigating the risk of post-dialytic hypokalemia, and potentially reduce the electrolyte-shift induced arrhythmias.
More arrhythmia data comes from a study in which patients had a cardiac loop recorder in situ (MiD trial, Roy-Chaudhary et al. Kidney Int, 2018). This study reported 1678 events (including 1461 bradycardia, 14 asystole, and one sustained ventricular tachycardia) in 44 patients. Atrial fibrillation was recorded in 41% of patients. These arrhythmias were highest in the 12 hours preceding dialysis, particularly the long interval, but only serum sodium and dialysate calcium (but not potassium) was found to be associated with clinically significant arrhythmias. Notwithstanding this, the loop recorders allow a mechanism for capturing data on the pathway of how serum and dialysate potassium might affect cardiac outcomes.
The Study
Methods
Study Design
The concept of introducing a potassium-binding resin with a 3 mEq/l potassium dialysate bath was tested in the ADAPT trial. This trial was a prospective, randomised, open-label crossover study. The authors considered the open-label design because of challenges associated with blinding patients’ dialysis solution, and due to the risk of improper potassium delivery, a blinded study was deemed unethical for safety reasons. The study design consisted of a 2-week screening period, followed by two 8-week crossover treatment phases. In addition, patients had a subcutaneous implantable loop recorder (ILR-LINQ device) inserted along the left parasternal border.
Source: Dr Sukhjinder Nijjer’s web page
Interventions
Individuals were randomised 1:1 into two groups:
Sequence A) Standard 2.0 K/2.5 Ca mEq/l (2 mmol/L K/1.25 mmol/L Ca) dialysate bath without SZC for 8 weeks followed by a 3.0 K/2.5 Ca dialysate bath along with oral SZC on non-dialysis days for 8 weeks
Sequence B) The groups then switched potassium baths and use of SZC for the second 8 weeks of the study
Figure 1 Charytan et al, Kidney Int, 2024
Patients with 3.0 K+ dialysate baths received SZC on off-dialysis days (4 days/ week). The starting dose was 5g/day (4/7), titrated weekly by 5 g up to 15 g, to maintain pre-dialysis K levels between 4-5.5 mEq/L. If K remained ≥ 5.5 after one week on 5g, the dose was increased to 10g for the next 7 days or as per investigator discretion. The maximum dose was 15 g/day.
SZC can transiently increase gastric pH, changing the absorption of co-administered drugs that exhibit pH-dependent solubility. Therefore, patients on other oral medications were advised to take other medications at least 2 hours before or 2 hours after SZC.
To evaluate compliance, patients were asked to return empty and unused SZC packets weekly during the 8-week 3K dialysate period. Pre-dialysis K levels after long dialysis served as an internal compliance check. The unused drug was recorded, and the remaining pockets were destroyed locally following the sponsor’s approval.
Eligibility
The inclusion and exclusion criteria are depicted in the infographic. It included only the adult population, with in-center, 3 times per week dialysis. Mentionable is that as per protocol, site PIs attempted to enroll at least 25% of their patients with a known history of post-dialysis intermittent (but not chronic) atrial fibrillation.
Endpoints
Primary efficacy endpoint:
Incidence of AF (defined as irregular heart rhythms and irregular R-R intervals in the absence of definable P waves for ≥ 2 min) with events occurring during treatment periods.
Secondary endpoints were efficacy and safety endpoints, detailed in the tables. Exploratory endpoints were not published in this paper.
‘Clinically significant cardiac arrhythmias’ (CSCA; labelled as clinically significant, but notably asymptomatic) were defined as: bradycardia ≤40 beats per minute [BPM] for ≥6 sec, ventricular tachycardia ≥130 BPM for ≥30 sec, asystole, and supraventricular tachycardia (SVT).
The CSCA were validated by the primary cardiologist blinded to the dialysate potassium content.
Funding
The trial was funded by AstraZeneca through an organization called NephroNet. The study was proposed, and the original protocol was drafted by the corresponding author (JAT). The approved, finalised protocol reflected modifications as suggested by AstraZeneca. Statistical analysis was performed by a blinded statistician at AstraZeneca. No third-party, or independent steering committee oversaw the trial. Few of the authors received consulting fees from AstraZeneca, while the first author declared receiving research funds from AstraZeneca and Medtronic. Editorial support (aka manuscript writing support) was provided by Carl V. Felton and Jess Fawcett, both of Core (a division of Prime, London, UK) funded by the sponsor.
Statistical considerations
To meet 90% power with two-sided alpha of 0.05, 40 subjects would need to be randomised into one of two treatment sequences in the 2x2 cross-over study. Accounting for this estimated 10% attrition rate, 44 subjects per treatment sequence, or 88 subjects in total will be randomised in the study.
The full analysis set included all randomised participants for efficacy analysis. Per protocol analysis included patients from full set analysis without major protocol deviations. For safety analyses, all the patients who received at least one dose of SZC were included.For primary outcomes a negative binomial model with two-sided was used, with alpha of 0.05, accounting for period and treatment effects. For secondary outcomes, the authors used separate models without multiplicity correction. Sensitivity analyses were conducted using multiple imputations for missing data. The authors didn’t plan for any interim analysis.
Results
Of the 148 participants screened for eligibility, 88 met the inclusion criteria for the trial, while 60 patients were excluded according to inclusion and exclusion criteria.
Figure S1 Charytan et al, Kidney Int, 2024
The mean age was 57 years and the majority were males. The mean pre-dialysis potassium level was 5.5 mEq/L. Few patients had a prior history of AF (8%) or CSCA (13 %).
Table 1. Charytan et al, Kidney Int, 2024
SZC dosing
The mean starting dose was 8.1g (SD 3.6) and the mean final dose was 10.2g (SD 4.0).
Figure 2. Charytan et al, Kidney Int, 2024
Starting and ending doses of SZC (Sodium Zirconium Cyclosilicate)
AF and CSCA Events
Frequency
There were 296 AF episodes noted in 9 participants, including 2 out of the 7 participants who had a prior history of AF. 6 participants in the 3.0K/SZC group recorded 123 events, whereas 7 participants in the 2.0K/no SZC group recorded 173 events. It was evident that the use of 3.0K/SZC significantly reduced the incidence of AF compared to the 2.0K/no SZC group.
Table 2. Charytan et al, Kidney Int, 2024.
Event counts and rates for AF and clinically significant cardiac arrhythmias.
There were 86 events of CSCA detected in 11 patients with 3.0K/SZC and 131 events in 13 patients with 2.0K/no SZC. The type of CSCA are shown below and included asystole, bradycardia, supraventricular, and ventricular arrhythmias. However for both AF and CSCA, the timing of these events (pre/post dialysis) is not reported.
From supplementary figure S2 Charytan et al, Kidney Int, 2024
Clinically significant cardiac arrhythmias, (a) overall and by type of arrhythmia: (b) asystole, (c) bradycardia, (d) supraventricular tachycardia, and (e) ventricular tachycardia
Timing
The number of AF events by participant, week, and treatment are as shown below. As noted in this figure, patients on 2K+/no SZC had numerically more episodes of atrial fibrillation as compared to those on 3K+/SZC. Not only were the absolute number of events higher in the 2K/ no SZC group, but the number of participants experiencing events was also higher. It is important to note only two patients (number 5 and 9) accounted for a majority of events.
Figure 3. Charytan et al, Kidney Int, 2024
AF events by participant, week, and treatment. The number inside the symbol represents the number of events for that participant during the week
Event duration
The mean duration of AF events was 1191 min/year for 3.0 K/SZC compared with 2804 min/year for 2.0 K/no SZC group. However, there were no significant differences in the mean proportion of follow-up time spent in AF between the two exposure groups.
Serum potassium measurements
The following figure shows the changes in pre-HD and post-HD serum potassium over time. 3.0K/SZC significantly reduced the chances of serum potassium being outside the optimal window; modeled odds ratio 0.27 (95% CI 0.21 to 0.35).
Figure 4. Charytan et al, Kidney Int, 2024
Serum potassium measurements outside the optimal window: (a) number of events by week, and (b) distribution of number of weeks with events.
Safety
The safety data revealed that:
The 3K/SZC group had lower incidence rates of post dialysis hypokalemia events, defined as serum K < 3.5 mmol/L (33) compared to the 2K/no SZC group (58).
No significant differences in the incidence of dialysis-related hypomagnesemia or hypophosphatemia were noted.
The interdialytic weight gain during the long interval was comparable in both groups (of interest since SZC does contain sodium).
The 3K/SZC group reported an increased rate of adverse events (43%) compared with the 2K/no SZC group (33%). Despite this, none of the adverse events were therapy related which led to dose interruption.
At least one serious adverse event (SAE) was experienced in 22 participants (15 for 3K/SZC, and 11 with 2K/no SZC). The most commonly reported SAEs were respiratory, thoracic, and mediastinal followed by metabolic and nutritional disorders.
Hyperkalemia was the most frequently reported SAE.
There were three cardiovascular SAEs during 3K/SZC treatment and four SAEs during 2K/no SZC
Discussion
Intradialytic potassium homeostasis can be maintained if the potassium flux is minimized. Utilizing higher concentrations of dialysate potassium can be an effective approach when combined with intestinal potassium binders- this can mitigate the risk of intra and post-dialytic hypokalemia. The recently approved patiromer sorbitex calcium and SZC can be utilised for this purpose.
The effects of SZC were first explored in the HARMONIZE trial (Kosiborod et al, JAMA, 2014). HARMONIZE was a phase 3, multicenter, randomised, double-blind, placebo-controlled trial evaluating zirconium cyclosilicate in outpatients with hyperkalemia (serum potassium ≥5.1 mEq/l). It recruited participants from 44 sites in the United States, Australia, and South Africa (March-August 2014). Zirconium silicate is a non-absorbable potassium binding agent. It is an inorganic crystalline cation exchanger with the capacity to bind both potassium and ammonium in the GI tract. The non-absorbable nature has been promoted as the key to minimising systemic side effects. The inclusion criteria of that study was simply a serum K of ≥ 5.1 on two occasions at least five minutes apart. No dialysis or end-stage renal disease patients were included. In the trial, 258 patients received 10g of zirconium cyclosilicate three times daily in the initial phase. In the subsequent 28 days, 237 patients achieving normokalemia (3.5-5.0 mEq/L) within the initial 48 hours were then randomised to receive zirconium cyclosilicate, 5g (n = 45 patients), 10g (n = 51), or 15g (n = 56), or placebo (n = 85) daily. The study revealed that among outpatients with hyperkalemia, open-label SZC reduced serum potassium to normal levels within 48 hours. All 3 doses of SZC resulted in lower potassium levels and a higher proportion of patients with normal potassium levels, for up to 28 days, compared to placebo.
Figure 3 of HARMONIZE trial (Kosiborod et al, JAMA, 2014).
The current study revealed that a strategy combining 3mEq/L K+ dialysate and SZC on non-dialysis days, which was tailored to maintain pre-dialysis serum K+ between 4 and 5.5 mEq/L, reduced the rates of AF and CSCA by half. Other significant findings were:
The odds of serum potassium outside the optimal window were lowered by more than 70% with 3K/SZC treatment due to a reduction in post-dialysis hypokalemia.
Incidences of hyperkalemia were infrequent in the 3.0K/SZC group.
Rates of post-dialysis hypomagnesemia and hypophosphatemia were not significantly different in both groups.
Use of SZC on non-dialysis days in the 3.0K/SZC group was well tolerated.
In the current study, the authors reported a reduction in the incidence of arrhythmias with the 3.0K/SZC strategy and concluded that the most important driver of arrhythmia in patients on dialysis was the flux of potassium during dialysis. However, the question remains whether this strategy can effectively mitigate the silent arrhythmias which pose a considerable threat to intradialytic outcomes. Moreover, serial intradialytic potassium testing is preferable to ensure that the intradialytic potassium homeostasis is restored and maintained using a higher concentration of potassium dialysate. This strategy should be practiced over longer periods of time to ensure its effectiveness.
Limitations:
This was a small and short trial, with surrogate and not clinical outcomes (don’t be fooled with the use of ‘clinically significant’ descriptor for asymptomatic arrhythmias)
The timings of the arrhythmias relative to the dialysis sessions were not assessed and they were of brief duration as well as asymptomatic
Difficult to generalize the results owing to a relatively young (mean age 57 years) study population and the fact that only 9 out of 88 participants had atrial fibrillation events.
It is worthwhile to note that the design of the ADAPT trial did not determine what proportion of the arrhythmia benefits were attributed secondary to the change of dialysate K+ of 2 to 3 mEq/L versus the benefit of SZC.
AF episodes > 6 min (because of being clinically significant) were in the initial protocol, but this was modified to on > 2 min without explanation. This may have lead to an increased reporting of less significant events
This is an investigator initiated but industry sponsored trial, taking place in non-academic centers, with involvement of the sponsor at all steps. We need to be careful about drawing definitive conclusions about the benefits of SZC and arrhythmia. Caveat emptor!
Conclusion
A strategy combining 3 mEq/L potassium dialysate and SZC on non-dialysis days, which was tailored to maintain pre-dialysis serum K+ between 4.0 and 5.5 mEq/L, reduced the rates of AF and CSCA by half. Longer and larger as well as blinded studies powered for clinical outcomes are needed to confirm these findings before we implement these into clinical practice routinely.
Summary Prepared by
Nikhil Elenjickal
Nephrologist, Bengaluru, India
Kajaree Giri
Nephrologist, Hyderabad, India
NSMC Interns Class of 2024
POD 1 Filtrate Firebolts
Reviewed by Brian Rifkin, Cristina Popa, and Swapnil Hiremath
Header Image created by AI, based on prompts by Brian Rifkin