Background
Hyperkalaemia is an electrolyte abnormality defined as a serum potassium (K
+) level above the normal physiological range of 3.5–5.0 mmol/L [
1,
2]. As renal secretion is the main route of potassium elimination, patients with renal and metabolic comorbidities are at an increased risk of hyperkalaemia [
3,
4]. Mild hyperkalaemia is usually asymptomatic, however even modestly raised serum K
+ (5.0–5.5 mmol/L) is associated with increased risk of adverse outcomes in patients with chronic kidney disease (CKD) [
5‐
7]. This risk profile may partly be compounded by the relationship between renin–angiotensin–aldosterone system inhibitors (RAASi) and hyperkalaemia. International and national guidelines recommend RAASi as first-line agents [
8‐
10] to delay CKD progression and to lower risks of kidney failure, cardiovascular disease and mortality [
11,
12]. However, despite these benefits, RAASi – which comprise angiotensin converting enzyme inhibitors (ACEi), angiotensin receptor blockers (ARBs) and mineralocorticoid receptor antagonists (MRAs) – can cause hyperkalaemia [
13]. Therefore, RAASi therapy is frequently downtitrated or discontinued in patients who experience hyperkalaemia, contributing to significant discrepancies between guideline recommendations and real-world practice regarding RAASi treatment [
13‐
16]. Failure to achieve guideline-recommended doses of RAASi has been associated with adverse outcomes [
8,
13,
17]. Thus, there is an established need for safe and effective treatment of hyperkalaemia, in particular to permit maintenance of RAASi therapy.
Although there is a well-defined treatment pathway for acute, life-threatening hyperkalaemia (K
+ > 7 mmol/L or with serious ECG changes [
18,
19]) in the inpatient setting, [
1] management regimes and thresholds for intervention for chronic hyperkalaemia are varied, typically region specific, and may suffer from significant limitations. In general, management of chronic hyperkalaemia comprises a combination of limiting K
+ intake through dietary modification, preventing K
+ retention by down-titration or discontinuation of medications such as RAASi, and the use of oral K
+ binders to facilitate K
+ excretion via the gastrointestinal route [
1]. Until recently, oral K
+ binders were limited to sodium polystyrene sulfonate (SPS), which is used in both the inpatient and outpatient setting in Sweden, and calcium polystyrene sulfonate (CPS), which is used in Norway, however the efficacies of SPS and CPS are uncertain and they are associated with adverse events (AEs) [
2,
20‐
22].
Novel oral K
+ binders that safely and effectively manage serum K
+ may be used to achieve guideline-directed doses of RAASi therapy, [
1,
2] which has significant potential to improve long-term outcomes in CKD patients [
1]. Sodium zirconium cyclosilicate (SZC) is a novel oral K
+ binder that has been shown in the ZS-003 and HARMONIZE trials [
23,
24] and two open-label studies, ZS-004E and ZS-005 [
25,
26] to significantly reduce serum K
+ in patients with hyperkalaemia and to subsequently maintain normokalaemia. SZC has been recommended by Norwegian Medicines Agency for patients with CKD who develop hyperkalaemia due to RAASi treatment and for patients with heart failure (HF) with serum K
+ ≥ 6.0 mmol/L, and by Dental and Pharmaceutical Benefits Agency for use in Sweden for adults with CKD stage 3–5, with or without chronic heart failure (HF), when resonium is not suitable; and for adults with chronic HF without comorbid CKD [
27‐
29]. The value of SZC at the population level, in terms of the long-term health and economic burden of hyperkalaemia and suboptimal RAASi therapy in CKD patients, has been limited due to modest treatment uptake, particularly in Sweden. Therefore, this study evaluated the cost effectiveness of SZC versus usual care – a combination of RAASi dose adjustments and intermittent SPS/CPS therapy – for the treatment of hyperkalaemia in patients with CKD in Norway and Sweden.
Discussion
This study modelled the clinical and economic impact of SZC versus usual care for the treatment of chronic hyperkalaemia in patients with CKD, to evaluate the cost effectiveness of introducing SZC treatment in Norway and Sweden. This analysis adapted an established simulation model [
30] that incorporates the natural history of CKD, serum K
+ fluctuations and changes in RAASi use to inform long-term health economic outcomes. The methods followed recommendations for health economic assessments to provide information for value-based decision making for healthcare services, and demonstrated that SZC is a cost-effective treatment for hyperkalaemia in both Norway and Sweden, compared to usual care, at a range of clinically plausible serum K
+ treatment thresholds.
Usual care for chronic hyperkalaemia comprises implementation of a low-potassium diet, down-titration or cessation of RAASi and intermittent SPS/CPS, but each of these has significant limitations. Recent KDIGO guidance on hyperkalaemia management in kidney disease highlighted that evidence supporting low-potassium diet is weak and that it may deprive patients of the benefits of a plant-rich diet [
1]. Similarly, downtitration or discontinuation of RAASi deprives patients of the benefits of RAAS inhibition and is associated with an increased risk of adverse outcomes [
8,
13,
17]. The evidence for the efficacy of SPS/CPS is modest; one small, short-term, single-blinded RCT has been published comparing SPS to CPS, [
34] however no comparison with placebo was performed and no studies have evaluated longer term efficacy or safety. Furthermore, SPS has been associated with serious AEs [
1] as well as poor palatability; consequently, ESC guidelines recommend not to use SPS in the medium or long term due to the risk of severe gastrointestinal side effects [
2,
20‐
22]. In contrast, SZC has been demonstrated in several trials to effectively treat hyperkalaemia and maintain normokalaemia, including in patients with CKD and those receiving dialysis, and has a favourable safety and tolerability profile [
23,
25,
26,
48]. Furthermore, the benefit of using novel oral K
+ binders to optimise RAASi use has been recognised by the ESC, [
2,
8] and this approach may also be beneficial for patients with CKD, as reflected in the 2020 UK Renal Association hyperkalaemia guidelines [
49].
In particular, therapies such as RAASi that slow the progression of CKD delay the need to initiate RRT. RRT, particularly dialysis, is associated with reduced quality of life, [
50] therefore strategies that delay RRT onset can provide substantial clinical benefit to patients with CKD. In addition to RAASi, other interventions aimed at slowing the progression of CKD are being introduced, such as the SGLT2 inhibitors. Although these have demonstrated efficacy in slowing eGFR decline, [
51] it is not anticipated that their introduction will impact on the present analysis. While slowing CKD progression is expected to reduce the overall incidence of hyperkalaemia caused by the increased risk of developing hyperkalaemia as renal function declines, the population considered in the present analysis are those who have already experienced a hyperkalaemia event and are thus at risk of recurrence.
The main limitations of this study relate to the necessary modelling assumptions required due to a lack of published evidence. In particular, as no trials have been conducted directly comparing the efficacy of SZC with either SPS or CPS, nor of SPS or CPS with placebo, we assumed the placebo arm of the HARMONIZE trial was representative of usual care during the maintenance phase. Intermittent SPS treatment is used on an outpatient basis in Sweden; it is unlikely that the data modelled fully capture this, therefore the efficacy may be underestimated in the usual care arm for the Swedish setting. Furthermore, a low-potassium diet was not mandated in the SZC trials, therefore it is uncertain to what extent the potential efficacy of such a diet may impact either the modelled SZC or usual care arms. However, as evidence supporting the efficacy of low-potassium diets is weak and adherence is challenging, our approach may be considered reflective of real-world practice. Additionally, patients in the SZC trials received treatment for concomitant acidosis as required, therefore the effect of this on hyperkalaemia is also implicitly captured in the modelling.
Where there was uncertainty in parameters, the base case sought to be conservative. For instance, the K
+ trajectories of usual care from Day 29 + were assumed to be equal to Days 15–28, although it is likely that K
+ levels would elevate as CKD progresses. Additionally, no evidence was found to allow a specific disutility of hyperkalaemia to be modelled, even though hyperkalaemia is associated with adverse outcomes [
5‐
7]. Another conservative assumption was that CKD patients with comorbid HF do not derive additional benefit on HF disease progression from maintaining RAASi.
An additional limitation is that this analysis did not consider patiromer, another novel oral K
+ binder. There is no head to head comparison between SZC and patiromer; while similar efficacy and safety profiles for normokalaemia maintenance are reported, [
25,
52] patiromer’s slower onset of action means it may not be suitable for rapid control of hyperkalaemia [
53]. There are additional benefits to patients and healthcare providers that were not captured in the cost-effectiveness analysis: SZC, unlike patiromer, can be stored at room temperature, and has a less restricted dosing window, which are meaningful in terms of facilitating the logistics of dispensing services and reducing the medication burden for patients with CKD. Additionally, SZC is licensed for use in patients receiving dialysis, and is particularly relevant for those dialysis patients who become anuric.
The HARMONIZE trial population was younger and contained a higher proportion of males compared to data published on a real-world Danish cohort of newly diagnosed CKD patients [
54]. This is a typical limitation of using clinical trial data. Furthermore, the HARMONIZE trial contained a higher proportion of patients with diabetes than the Danish cohort, [
23,
54] however, the trial is still expected to be broadly generalisable to patients and clinical practice in Sweden and Norway, because patients in clinical practice are expected to have a longer history of CKD compared to those enrolled in the Danish study, and the number of comorbidities, including diabetes, and the risk of hyperkalaemia, are expected to increase with longer history of CKD [
55,
56].
Therefore, despite these limitations, we consider the analysis to be broadly generalisable to Swedish and Norwegian patients and clinical practice. Another strength was that no extrapolation of efficacy was required; all efficacy data derived from the trial follow-up period. Furthermore, SZC remained cost effective in all clinically-plausible scenarios examined, and the results of the sensitivity analyses indicated that the cost-effectiveness results were robust to assumptions. The highest impact parameter was the threshold of serum K
+ for initiating active treatment. This was explicitly explored in a scenario analysis where the serum K
+ threshold for treatment was changed to ≥ 5.1 mmol/L and the K
+ trajectories were derived from the full HARMONIZE, ZS-004E and ZS-005 data sets [
23,
25,
26]. Current clinical practice in Sweden and Norway is to initiate active treatment at a serum K
+ threshold of 5.5 mmol/L, as in the base case. This choice of threshold represents a clinician-led risk–benefit analysis of mild hyperkalaemia against the AEs, tolerability and poor palatability of SPS/CPS. However, SZC provides an effective alternative with a proven safety and tolerability profile, which may lead to a re-evaluation of the treatment threshold, set against the increase in risks of adverse outcomes associated with serum K
+ ≥ 5.1 mmol/L. The demonstration of the cost effectiveness of SZC at the lower threshold provides further evidence in support of this.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.