Cost-effectiveness analysis of sodium zirconium cyclosilicate for hyperkalemia among patients with chronic kidney disease or heart failure in Kuwait

Abstract

Introduction

Our model was conducted from Kuwaiti payer’s perspective to provide evidence on the cost-effectiveness of Sodium zirconium cyclosilicate (SZC) versus patiromer to correct and maintain serum potassium (K+) in combination with renin‐angiotensin‐aldosterone system inhibitors (RAASis) with different dose titration in patients with chronic kidney disease/heart failure (CKD/HF) with/without renal replacement therapy (RRT).

Methodology

The model was developed as a patient-level, fixed-time increment stochastic simulation to simulate the complexity of disease, including multiple coexisting and competing conditional risks. This model was established to compare SZC versus patiromer as a treatment for hyperkalemia (HK) among adult populations with underlying conditions of advanced CKD stages 3a–5 or HF to correct and maintain serum K + over a lifetime horizon. The clinical outcomes of SZC and patiromer were demonstrated through arm-specific K + trajectories extracted from the HARMONIZE trial and OPAL-HK trial, respectively. The utility data was captured from different studies. Direct medical cost was captured from local data from Kuwaiti hospitals. Sensitivity analyses were conducted to assess the uncertainty in the model.

Results

Within different scenarios of CKD/HF, SZC was a cost-saving option, with/without RRT, whether one-off administration or repeated administration, except for one-off treatment administration among the HF cohort, which generated an incremental cost effectiveness ratio of KWD 331/quality adjusted life year (QALY). The incremental QALY of SZC ranged from 0.007 to 0.202. In addition, the savings observed with SZC fall within a range of KWD −60 to KWD −1,235 at serum K+ ≥ 5.1 mmol/L.

Conclusion

The evidence generated by our model recommends the inclusion of SZC as a treatment option to correct HK and maintain normal serum K + level for CKD/HF patients within the Kuwaiti healthcare system. The costs saved from reducing frequent HK episodes, RAASis discontinuation/down titration, major cardiovascular events, and hospitalization offset the drug acquisition cost of SZC.

Keywords:

• Chronic kidney disease
• heart failure
• cost-effectiveness analysis
• sodium zirconium cyclosilicate

JEL Classification Codes:

• C52
• C5
• C
• E47
• E4
• E

Introduction

Hyperkalaemia (HK) is a life-threatening condition defined as an electrolyte abnormality of serum or plasma potassium (K+) levels above the upper limit of its normal physiological range^1,2. A major risk factor for HK is chronic kidney disease (CKD) due to a reduction in renal excretion of K+, especially if it is accompanied by heart failure (HF)^2,3. In addition to CKD/HF, the administration of renin‐angiotensin‐aldosterone system inhibitors (RAASis), which are the cornerstone to prevent the progression of CKD/HF, is known to increase the risk of HK⁴. HK associated with RAASis can lead to treatment discontinuation or downtitration, which causes suboptimal management of the disease and the loss of benefits of these medications⁵.

There is a lack of data regarding the prevalence of HK in CKD/HF patients in Kuwait⁶. Between 1976 and 2021, the pooled prevalence of HK (according to any threshold) among the adult population was 6.3%⁷, with an estimated incidence of 2.8 cases per 100 person-years⁷. A study by Kanda et al. addressed the importance of HK treatment, as it was associated with frequent hospitalization that led to long-term health care costs⁸. Additionally, resource utilization and health care costs remained high for HK patients after 12 months⁸.

Determination of serum K + abnormalities is very important, as they may cause life-threatening outcomes, such as cardiac arrhythmia, muscle paralysis and sudden cardiac death^1,9. Furthermore, HK is associated with increases in mortality rates, hospitalization and emergency department (ED) visits¹⁰. Among patients with severe HK (K + levels ≥ 6.5 mEq/L), the in-hospital mortality rate was 30.7%¹¹. A study conducted in the United Kingdom (UK) demonstrated that patients with severe index HK had a higher rate of adverse outcomes, such as arrhythmia and death¹². In the United States (US), compared to CKD/HF patients without HK, CKD/HF patients with HK incurred higher all-cause total health care costs within 30 days ($5,553) and within 1 year ($24,133)¹³.

A large retrospective study conducted in Canada reflected that one single episode of HK (serum K+ ≥ 5.0 mmol/l) led to a financial burden over the short and long term¹⁴. The study concluded that HK is linked with adverse clinical outcomes such as hospitalizations, intensive care unit (ICU) admissions, all-cause mortality, and cardiovascular (CV) events¹⁴. The hazard ratios (HRs) of CV events and all-cause mortality were 1.20 (95% confidence interval (CI) = 1.14–1.26) and 1.15 (CI = 1.13–1.18), respectively¹⁴.

Luo et al. assessed the relationship between serum K + levels and major adverse CV events (MACEs) among CKD patients¹⁵. Across subgroups of estimated glomerular filtration rate (eGFR), there was a consistent relation (U-shaped association) between K + level and MACE¹⁵. In addition, Luo et al. demonstrated a statistically significant relationship between high serum K + levels and an increased rate of RAASis discontinuation¹⁵. Therefore, control of K + levels through safe and effective pharmacotherapies is an important aspect to reduce the risks of hospitalization, MACE, mortality and RAASis discontinuation or downtitration^5,16.

Sodium zirconium cyclosilicate (SZC; Lokelma®, by AstraZeneca), a novel oral K + binder, causes a significant reduction in serum K + in patients with HK and thereafter maintains normokalaemia. The effect of SZC has been shown in two open label studies of ZS-005¹⁷ and ZS-004E¹⁷ and the HARMONIZE trial (ZS-004)¹⁸.

ZS-005, a 12-month international, phase-3, multicentre, single-arm trial, evaluated the efficacy and safety of oral SZC to restore and maintain normokalaemia among an adult population with serum K+ ≥ 5.1 mmol/L with different comorbidities, such as CKD or HF, which needed management by different concomitant medications, such as RAASis¹⁹. Within ZS-005, 99% of patients reached rapid normokalaemia within 24–72 h during the correction phase. For those who entered the maintenance phase, SZC made long-term maintenance of normokalaemia up to 12 months¹⁹. In addition, SZC-related normokalaemia did not require any substantial change in RAASi use¹⁹.

Objective

Data regarding the value of SZC with respect to the long-term health and economic burden of HK among CKD/HF patients with RAASis have been limited, particularly in Kuwait. Thus, our economic model was conducted from the Kuwaiti payer’s perspective (the healthcare system) to provide evidence on the cost-effectiveness of SZC versus patiromer to correct and maintain serum K + in combination with RAASis with different dose titration in patients with CKD/HF with or without RRT over a lifetime horizon.

Methodology

Model overview

The model was developed as a patient-level, fixed-time increment stochastic simulation (Markov individual simulation) to simulate the complexity of disease, including multiple coexisting and competing conditional risks. The model has been developed in Microsoft Excel for Windows, and is compatible with Excel 2007 or later. The model’s core calculations are undertaken within Visual Basic for Applications. The model simulates the natural history of disease for patients with advanced CKD and/or HF, using standard approaches to classify disease severity and predict disease progression. Treatment-specific K + trajectories were modelled at the patient-level, and linked to the risk of fatal and non-fatal events, including changes in RAASi therapy. This model was established to compare SZC (10 g and 5 g) versus patiromer as a treatment for HK among adult populations with underlying conditions of advanced CKD stages 3a–5 (eGFR = ≥0–<60 mL/min/1.73 m²) or HF (New York Heart Association; NYHA functional class I, II, III or IV). For SZC or patiromer, the correction phase of initial acute HK lasts for three days in the hospital, while the maintenance phase of treatment lasts for 52 weeks outside the hospital.

All patients in the HF/CKD cohort were on RAASi therapy at model initiation, as this is a recommended treatment strategy for this cohort. At baseline, the simulated patients were 64 years old, 86 kg with an eGFR of 47 ml/min/1.73 m², which was consistent with the characteristics of patients from ZS-005¹⁹.

To ensure stable point estimation, the simulated cohort was 30,000 patients either for HF or CKD. We assumed that the proportions of patients in each NYHA class I, II, III and IV at baseline were 0.4, 0.3, 0.2 and 0.1, respectively.

Owing to the lack of consensus in real-life practice about the appropriate K + threshold to initiate treatment, our model simulated two scenarios of HK events (both initial and recurrent episodes) at a serum K + threshold of ≥ 5.1 and ≥5.5 mmol/L. In case of treatment discontinuation, patients in both arms were either administered sodium polystyrene sulfonate (SPS) alone (one-off administration of SZC or patiromer) or SPS until reinitiation of treatments in both arms according to modelled K + level (SZC or patiromer repeat). Our model simulated patient transition between different CKD and HF health states. As simulated patients progress through the model, assessments are made for fatal and non-fatal events. The patients in our model may experience nonfatal events such as MACEs, hospitalization, acute HK if the serum K + level exceeds a defined threshold (either ≥ 5.1 or ≥5.5 mmol/L) or a change in RAASi use. Hospitalization, changes in RAASi use, MACE and mortality were dependent on serum K + levels. Thus, we assumed that those patients have totally non-fatal events (no acute death from these events) and the only way for death is the progression to dialysis. Figure 1 shows health states and events in our model.

Figure 1. The health states and events in the model. Abbreviations. HF, heart failure; NYHA, New York Heart Association; MACE, Major adverse cardiovascular events; AEs, adverse events; RAASi, renin–angiotensin–aldosterone inhibitors; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; RRT, renal replacement therapy; ESRD, end stage renal disease

This manuscript was reported in accordance with the International Society for Health Economics and Outcomes Research (ISPOR) Good Reporting Practices Task Force: Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement²⁰. All inputs were captured from a literature review of all clinical trials conducted on the available treatments of HK among CKD and HF patients, validated by our expert panel (two cardiologists, two nephrologists and two health economists). Expert insights included the current local clinical practice and treatment patterns of these patients within Kuwaiti health care settings. Table S1 includes the survey conducted with the expert panel.

The outputs measured in our study were the costs, quality-adjusted life-years (QALYs), life years (LYs) and incremental cost-effectiveness ratio (ICER). ICER was reported by computing the incremental costs measured in Kuwaiti Dinar (KWD) an defectiveness then dividing it by the incremental health benefits calculated as QALYs of SZC versus patiromer to correct HK during the correction phase and maintain normal serum K + levels during the maintenance phase as per the following equation: $$\text{Cost of SZC}–\text{Cost of patiromer}/\text{Effectiveness of SZC}–\text{Effectiveness of patiromer}$$

The cycle length in our model was four weeks (28 d). To simulate changes in serum K + levels during the correction phase, the first four-week period was broken into shorter cycle lengths as follows: cycles 1, 2 and 3 represented Day 1, Day 2 and Day 3, respectively; cycle 4 represented Days 4–14; and cycle 5 represented Days 15–28. From cycle 6, the cycle length was 4 weeks.

A half-cycle correction was applied based on ISPOR modelling good research practices²¹ to adjust the distribution of costs and effects that occurred throughout the cycle. The lifetime horizon (13 years) of both scenarios was considered long enough to enable capturing the long-term survival benefits of improved management of serum K+. Costs and health-related outcomes were discounted at a rate of 3.5% in the base case (mean value) in accordance with the National Institute for Health and Care Excellence (NICE) recommendations for discounting²².

Clinical inputs

SZC and patiromer K + trajectories were modelled at the patient-level, and linked to the risk of fatal and non-fatal events, including changes in RAASi therapy, extracted from the HARMONIZE trial (ZS-004) and OPAL-HK trial, respectively^18,23. These patient-specific K + profiles were simulated using mixed-effects regression models. Fixed effects model parameters captured overall K + trends at the population-level (represent the time-varying, population-averaged mean level of serum K+), while the inclusion of random effects terms allowed individual patients to exhibit a unique, fluctuating K + profile that may deviate from the population average (represent patient-specific mean K + levels that may be systematically higher or lower than the population-averaged mean level). The inclusion of patient-specific random effects increases the accuracy of the models, ensuring that key statistical inputs were satisfied and inferences obtained from the models are valid. For each simulated patient, a new K + value was sampled in each modelled cycle. To reflect the design of the ZS-004 and ZS-005 trials, K + values were sampled daily in the acute phase (days 0–3). During the maintenance phase (days 4+), K + values were sampled once in days 4–14, once in days 15–28, and once in each 28-day cycle thereafter.

HARMONIZE, a phase-3, randomized, multicentre controlled trial, evaluated the safety and efficacy of SZC among 258 patients with HK (K+ ≥5.1 mEq/L) against placebo¹⁸. Within HARMONIZE, 84% and 98% of the patients achieved normokalaemia after administration of 10 g SZC 3 times daily within 24 h and 48 h, respectively¹⁸. During Days 8–29, SZC caused a significant reduction in serum K + levels compared to placebo¹⁸.

OPAL-HK is a single-blind study that evaluated the efficacy and safety of patiromer to manage HK among 237 patients with CKD who received RAASis²³. Within OPAL-HK, 42% of the overall cohort had HF. On Day 3 of the initial treatment phase, the mean change in serum K + by the patient was −1.01 ± 0.03 mmol/litre, while 76% of the patients reached normokalemia after 4 weeks²³. The justification behind utilizing patient outcomes among the HF/CKD cohort from OPAL-HK over PEARL-HF was due to the higher number of enrolled patients (237 vs. 120) and longer duration of assessment (8 weeks vs. 4 weeks)^23,24.

For the CKD/HF cohort, all patients were administered RAASis at model initiation. Discontinuation or dose adjustment with RAAS is occurred when serum K+ ≥ 5.1 mmol/L²⁵. The proportion of patients who returned to the optimal RAASi dose was 50% (0.497) within 42 (± 4.2) weeks¹⁵.

HF cohort

The monthly transition probabilities of HF patients, from each NYHA class to another, were extracted from a long-term cost-effectiveness study that evaluated treatment with cardiac resynchronization therapy among patients with NYHA class III/IV²⁶.

All-cause mortality of HF patients, MACE and hospitalization were dependent and modelled according to serum K + intervals. Additionally, hospitalization was adjusted by RAASi use. The HR of all-cause mortality in HF patients was extracted from a study that evaluated short-term (90-day follow-up) mortality risk among 2,596 patients following myocardial infarction (MI)²⁷. The monthly probability of all-cause hospitalization was calculated by the Clinical Practice Research Datalink (CPRD) risk equation²⁸. The incidence rate ratio (IRR) of MACE for the HF cohort was assumed to be equal to that of the CKD cohort based on our expert panel.

CKD cohort

The rate of eGFR decline is dependent on RAASi use. The annual eGFR decline for patients with RAASis was extracted from a study that evaluated the efficacy of irbesartan to delay the progression of eGFR among 579 patients against 569 in the placebo group²⁹.

The annual rates of all-cause mortality, hospitalization and MACE were modelled according to CKD stages³⁰, adjusted by serum K + level¹⁵ and RAASi use³¹. We assumed that the impact of the submaximum dose of RAASis was 50% of the impact of the optimal dose of RAASis, with equal SE. Table 1 includes all clinical inputs of the model.

Table 1. The clinical parameters used in the model.

Parameters					References
Change in K level in SZC arm (captured from matched ZS-004)					[19]
Change in K level at days:	Day 0–3	Day 4 – 14	Day 15 – 28	Longer-term efficacy; Day +29
Fixed: Intercept	5.600	4.565	4.658	4.710
Random: Patient	0.225	0.213
Random: Observation	0.228	0.234
Change in K level in Patiromer arm (captured from Matched OPAL-HK)					[23]
Change in K level at days:	Day 0–3	Day 4 – 14	Day 15 – 28	Longer-term efficacy; Day +29
Fixed: Intercept	5.600	5.077	4.880	4.770
Random: Patient	0.237	0.216
Random: Observation	0.240	0.238
Change in K level in Patiromer arm (captured from Crude OPAL-HK)					[23]
Change in K level at days:	Day 0–3	Day 4 – 14	Day 15 – 28	Longer-term efficacy; Day +29
Fixed: Intercept	5.574	4.807	4.550	4.560
Random: Patient	0.322	0.429
Random: Observation	0.326	0.397
Transition probabilities among HF cohort
From\To Mean (SE)	NYHA I	NYHA II	NYHA III	NYHA IV	References
NHYA I	0.7956 (0.1)	0.1245 (0.1)	0.0738 (0.1)	0.0061 (0.1)	[26]
NYHA II	0.071 (0.1)	0.8448 (0.1)	0.0765 (0.1)	0.0077 (0.1)
NYHA III	0.0047 (0.1)	0.0893 (0.1)	0.8845 (0.1)	0.0216 (0.1)
NYHA IV	0	0.1064 (0.1)	0.1064 (0.1)	0.7872 (0.1)
K + Profile				Monthly probability of hospitalization Mean (SE)
RAASi: K + < 5.5				0.067 (0.004)	[48]
RAASi: K+ < 6.0				0.083 (0.009)
RAASi: K+ ≥ 6.0				0.117 (0.022)
No RAASi: K + < 5.5				0.105 (0.005)
No RAASi: K+ < 6.0				0.105 (0.010)
No RAASi: K+ ≥ 6.0				0.110 (0.021)
Sub RAASi: K + < 5.5				0.081 (0.005)
Sub RAASi: K+ < 6.0				0.091 (0.009)
Sub RAASi: K+ ≥ 6.0				0.114 (0.021)
All-cause mortality Mean (SE)
HR of Mortality	K + subgroup							[27]
	<3.5	3.5-3.8	3.9-4.2	4.3-4.5	4.6-5.0	5.1-5.5	>5.5
	2.19 (0.59)	1.91 (0.40)	1	1	1.47 (0.27)	2.28 (0.55)	6.6 (1.70)
Annual mean of eGFR decline (SE) in CKD cohort
CKD on RAASi				2.34 (0.023)	[28]
Annual event rate according to CKD stages (SE)
Event	CKD subgroups				[29]
	3a	3b	4	5
MACE	0.0365 (0.0012)	0.1129 (0.0012)	0.218 (0.0024)	0.366 (0.0048)
Hospitalization	0.1722 (0.0045)	0.4526 (0.0067)	0.8675 (0.0090)	1.4461 (0.0090)
All-cause mortality	0.0108 (0.0004)	0.0476 (0.0007)	0.1136 (0.0018)	0.1414 (0.0031)
IRR for events by K + subgroup mean (SE)
Event	K + subgroup							[15]
	<3.5	3.5-3.9	4-4.4	4.5-4.9	5-5.4	5.5-5.9	≥6
MACE	1.89 (0.09)	1.27 (0.04)	1.04 (0.01)	1	1.01 (0.02)	1.12 (0.04)	1.88 (0.12)
All-cause mortality	3.05 (0.29)	1.49 (0.08)	1.06 (0.04)	1	1.14 (0.06)	1.6 (0.13)	3.31 (0.46)
Hospitalization eGFR <30	1.93 (0.45)	1.65 (0.22)	0.93 (0.09)	1	1	1.34 (0.18)	3.65 (0.58)
Hospitalization eGFR 30-40	1.77 (0.34)	1.35 (0.16)	0.99 (0.08)	1	(0.10)	1.07 (0.14)	1.82 (0.44)
Hospitalization eGFR 40-50	2.24 (0.38)	1.23 (0.13)	1.08 (0.07)	1	0.96 (0.09)	1.23 (0.18)	1.91 (0.56)
Hospitalization eGFR 50-60	2.06 (0.27)	1.13 (0.08)	1.01 (0.05	1	1.07 (0.09)	0.81 (0.11)	1.07 (0.30)
Event risk by RAASi use mean (SE)
OR RAASi vs no RAASi: mortality	0.870 (0.069)	[30]
OR Sub RAASi vs no RAASi: mortality	0.935 (0.069)	Assumption
OR RAASi vs no RAASi: MACE	0.820(0.054)	[30]
OR Sub RAASi vs no RAASi: MACE	0.910(0.054)	Assumption

Abbreviations. CKD, Chronic kidney disease; eGFR, estimated glomerular filtration rate; HR, Hazard ratio; HF, Heart failure; MACE, Major adverse cardiovascular events; OR, Odds ratio; RAASi, Renin‐angiotensin‐aldosterone system inhibitors; SZC, Sodium zirconium cyclosilicate; CPRD, Clinical Practice Research Datalink; IRR, Incidence rate ratio; NYHA, New York Heart Association.

Outcomes

The utility values applied in our model are health-state dependent, not a treatment-specific utility. Among the CKD cohort, the utility of CKD stages 3a up to 4 was gathered through time trade-off (TTO) questionnaires to 205 patients with CKD that assessed the health-related quality of life (HRQoL) values among CKD patients³². The utility of CKD stage 5, dialysis and renal transplant status were extracted from a study that assessed HRQoL among renal failure patients by different methodologies, such as the EQ-5D index (applied in our model), SF-36 and the kidney disease Quality of Life Questionnaire (KDQOL)³³. Utility of haemodialysis was applied, as it is the most common dialysis type in Kuwait.

For the HF cohort, the utility of health states (NYHA class I–IV) was extracted from a study designed to assess utilities of different NYHA classifications by EuroQol-5D from 1,395 patients³⁴. The impact of significant treatment-related AEs and CKD/HF-related events on HRQoL was assessed by utility decrements and included in outcome measurements.

Our model considered the disutility of CKD/HF-related events such as hospitalization, dialysis complications, and MACE. The disutility of hospitalization was extracted from a study that assessed the utility of health states for NYHA class I–IV³⁴. The utility decrement of dialysis complications was collected from a cost–utility analysis comparing different modalities of dialysis³⁵. The decrement of MACE was collected from the economic evaluation of two competing strategies of either coronary angiography alone or in conjunction with fractional flow reserve measurement³⁶.

The SZC-related AE probabilities were extracted from 746 patients in the maintenance phase of ZS-005¹⁹, while AE probabilities within the patiromer arm were extracted from the OPAL-HK study of 298 patients administered patiromer (243 patients in the initial phase and 55 patients in the maintenance phase)²³. The disutility of AEs used in our model was extracted from different studies in the literature^37–39. The disutility of hypertension, constipation, hypomagnesemia, anaemia and urinary tract infection was extracted from a community-based catalogue of EQ-5D index scores that assessed different chronic conditions³⁷. The disutility of oedema was assumed to be equal to hypertension, while the disutility of hypokalaemia was assumed to be zero. The disutility of diarrhoea and nausea was extracted from two different studies published in the literature^38,39.

The utility decrements with SZC or patiromer were applied throughout the treatment duration. Therefore, the disutility was scaled to simulate the effect of each AE that would last 28 d within each year of treatment. Table 2 includes all utilities and disutilities used in the model.

Table 2. The utility inputs in the model.

	Base case	Low value	High value	References
Utilities:
CKD 3a	0.870	0.783	0.957	32
CKD 3b	0.870	0.783	0.957
CKD 4	0.850	0.765	0.935
CKD 5	0.570	0.513	0.627	32
Dialysis	0.44	0.396	0.484
Transplant	0.710	0.639	0.781
NYHA I	0.855	0.7695	0.9405	33
NYHA II	0.771	0.6939	0.8481
NHYA III	0.673	0.6057	0.7403
NYHA IV	0.532	0.4788	0.5852
Disutility
Hospitalization	−0.024	−0.0216	−0.0264	32
Dialysis complication	−0.02	−0.018	−0.022	34
MACE	−0.05	−0.045	−0.055	35
Edema (generalized and peripheral)	−0.0029	−0.00261	−0.00319	Assumption
Worsening hypertension	−0.0029	−0.00261	−0.00319	36
Constipation	−0.0056	−0.00504	−0.00616
Diarrhea	−0.0008	−0.00072	−0.00088	37
Nausea	−0.0037	−0.00333	−0.00407	38
Hypomagnesemia	−0.00073	−0.000657	−0.000803	36
Anemia	−0.0015	−0.00135	−0.00165
UTI	−0.0004	−0.00036	−0.00044

Abbreviations. UTI, Urinary tract infection; CKD, Chronic kidney disease; MACE, Major adverse cardiovascular events; NYHA, New York Heart Association.

Costs

Direct medical costs only were considered. The following direct medical costs were included: drug acquisition costs during the correction and maintenance phase, concomitant medications to manage CKD/HF, and monitoring and AE management costs. The frequency of resource use during routine follow-up was informed by our expert panel and based on Kuwaiti practices within local health care settings. Only significant treatment-related AEs were considered. AEs costs were calculated based on the average resources used and length of hospital stay to treat a single episode if needed. All unit costs were obtained from Jaber Al-Ahmed Armed Forces Hospital – Kuwait Ministry of Defence. Table 3 includes all the unit costs used in our model.

Table 3. Unit costs used in our model.

Cost parameters (KWD)
Intervention	Base case	Low value	High value	References
SZC 10 g	KWD 3.75	3.375	4.125	Jaber Al-Ahmed Armed Forces Hospital - Kuwait Ministry of Defense.
SZC 5 g	KWD 2.62	2.35	2.882
SPS 454 g	KWD 24.44	21.996	26.884
Patiromer (16.8 g − 8.4 g)	KWD 2.97	2.673	3.267
Furosemide 40 mg tablet	KWD 0.07	0.063	0.077
Furosemide 20 mg ampoule	KWD 0.18	0.162	0.198
Isosorbide mononitrate tab 60 mg	KWD 0.06	0.054	0.066
Amlodipine 5 mg	KWD 0.16	0.1413	0.1727
Bisacodyl 5 mg	KWD 0.01	0.0117	0.0143
Loperamide 2 mg	KWD 0.08	0.0675	0.0825
Metoclopramide 10 mg	KWD 0.02	0.0198	0.0242
Serum Mg level	KWD 1.00	0.9	1.1
Magnesium oxide 400 mg	KWD 0.03	0.0243	0.0297
Esomeprazole 20 mg	KWD 0.17	0.153	0.187
Epoetin Alpha 4,000 iu/ml	KWD 12.80	11.52	14.08
Nitrofurantoin	KWD 0.07	0.0648	0.0792
Amoxicillin/clavulanate 1 g	KWD 0.53	0.477	0.583
Physician visit	KWD 10.00	9	11
Outpatient visit	KWD 10.00	9	11
ER visit	KWD 10.00	9	11
Inpatient day (regular room)	KWD 10.00	9	11
Inpatient day (ICU room)	KWD 30.00	27	33
Serum creatinine	KWD 0.50	0.45	0.55
Urea	KWD 0.50	0.45	0.55
Serum K+	KWD 0.50	0.45	0.55
ECG	KWD 10.00	9	11
Regular insulin 100 iu/ml-3 ml	KWD 1.40	1.26	1.54
D50W (50% dextrose solution) 1,000 ml	KWD 0.78	0.702	0.858
Calcium gluconate iv. 10% w/v 10 ml	KWD 0.07	0.0657	0.0803	Jaber Al-Ahmed Armed Forces Hospital - Kuwait Ministry of Defense.
Salbutamol nebulizer2.5 mg	KWD 0.10	0.09	0.11
Ramipril 2.5 mg	KWD 0.24	0.216	0.264
Ramipril 5 mg	KWD 0.25	0.225	0.275
Ramipril 10 mg	KWD 0.26	0.234	0.286
Lisinopril 5 mg	KWD 0.03	0.027	0.033
Lisinopril 10 mg	KWD 0.06	0.054	0.066
Lisinopril 20 mg	KWD 0.10	0.09	0.11
Telmisartan 40 mg	KWD 0.13	0.117	0.143
Telmisartan 80 mg	KWD 0.19	0.171	0.209
Valsartan 80 mg	KWD 0.09	0.081	0.099
Valsartan 160 mg	KWD 0.10	0.09	0.11
Spironolactone 25 mg	KWD 0.06	0.054	0.066
Spironolactone 100 mg	KWD 0.17	0.153	0.187
Eplerenone 25 mg	KWD 0.80	0.72	0.88
Eplerenone 50 mg	KWD 0.80	0.72	0.88
CBC	KWD 2.00	1.8	2.2
Darbepoetin alfa 100 mcg/0.5 ml	KWD 76.77	69.093	84.447
Sevelamer 800 mg	KWD 0.27	0.243	0.297
Calcium carbonate + vit. D tablet	KWD 0.05	0.045	0.055
Normal saline	KWD 0.51	0.459	0.56
Carvedilol 6.25 mg	KWD 0.06	0.054	0.066
Empagliflozin 10 mg	KWD 0.40	0.36	0.44
Hemodialysis session	KWD 18.00	16.2	19.8
Heparin 5,000 iu/ml −5 ml	KWD 0.04	0.0333	0.0407
LFT	KWD 4.50	4.05	4.95
Lipid profile	KWD 5.00	4.5	5.5
Prednisone 5 mg	KWD 0.02	0.018	0.022
MPA 250 mg	KWD 0.24	0.216	0.264
MPA 500 mg	KWD 0.49	0.441	0.539
Cyclosporine 100 mg	KWD 1.30	1.17	1.43
Tacrolimus 1 mg	KWD 1.90	1.71	2.09
Trough level	KWD 10.00	9	11
Kidney transplantation cost	KWD 75.00	67.5	82.5
ECHO	KWD 40.00	36	44
Myoview stress test	KWD 50.00	45	55
Pulmonary function testing	KWD 30.00	27	33
Abdominal and pelvic ultrasonography	KWD 40.00	36	44
Upper GIT endoscopy	KWD 50.00	45	55
ABO	KWD 5.00	4.5	5.5
Virology scan	KWD 61.00	54.9	67.1

Abbreviations. SZC, Sodium zirconium cyclosilicate; SPS, Sodium polystyrene sulfonate; ER, emergency room; ECG, electrocardiogram; D50W, 50% dextrose solution; CBC, complete blood count; LFT, Liver function test; ECHO, echocardiogram; ABO, Blood Grouping System; MPA, Mycophenolic acid.

Treatments

HF/CKD cohort

All treatment lines options and doses used in our model are mentioned in Table S2. The background therapy for HF patients (NYHA class I, II, III and IV) was a quadruple therapy of angiotensin-converting enzyme inhibitor (ACEI)/angiotensin receptor blocker (ARB), beta-blocker (carvedilol), mineralocorticoid receptor antagonist (MRA), and sodium-glucose cotransporter-2 inhibitor (SGLT2i; empagliflozin).

The most common MACE was hospitalization due to HF, which needed a stay of the first 2 days in the ICU and then a 5-day ward with administration of nitrates and iv. diuretics. Hospitalization among HF/CKD patients leads to the development of acute kidney injury (AKI). Therefore, patients were admitted to the hospital for 4 days in a regular room.

CKD patients (3a to 5 stage) received mainly monotherapy with ACEIs or ARBs. Those CKD patients and those on dialysis needed further medications to assure case control as sevelamer, calcium carbonate + vitamin D 400 IU tablet and darbepoetin alfa. Renal replacement therapy (RRT) was considered in our model if the patients needed it. In Kuwait, the most common RRT used is a 4-h haemodialysis session (3 times/week) with concomitant administration of heparin. Monthly monitoring by complete blood count (CBC), urea, serum creatinine, lipid profile and liver function test (LFT) was performed. During dialysis, the most common complication was hypotension, which needed the administration of 250–500 ml iv. normal saline.

When renal transplantation is needed, the following assessments are needed: echocardiogram (ECHO), Myoview stress test, pulmonary function testing, abdominal and pelvic ultrasonography, upper gastrointestinal tract (GIT) endoscopy, virology scan and blood grouping system (ABO). Additionally, administration of cyclosporine and tacrolimus is needed pre/postoperative. After transplantation, patients are administered prednisone (10 mg) and mycophenolic acid (MPA), with monthly monitoring of CBC, serum creatinine, urea and tacrolimus serum trough levels.

Initial acute HK event

The severity of acute HK was categorized according to serum K + fluctuations into low (K+ ≥ 5.1 or 5.5 mmol/L) and moderate to severe (K+ ≥ 6.0 mmol/L). Therefore, the management of acute HK events varies according to serum K + levels.

For the management of acute HK with low severity, assessment of serum K + and kidney function test (KFT) were performed. In cases of acute moderate to severe HK events, patients would require initial assessment by KFT, serum K + and electrocardiography (ECG) in the emergency department. Then, patients are hospitalized for 3 d with SZC, while 4 d with patiromer or SPS due to rapid correction of HK with SZC proven by ZS-005¹⁹ and HARMONIZE trial (ZS-004)¹⁸ with administration of temporizing measurement as; iv. calcium gluconate, iv. regular insulin with iv. 50% dextrose solution (D50W) and salbutamol nebulizer.

Recurrent HK event

Episodes of recurrent HK could occur due to treatment discontinuation with SZC or patiromer (0.26 and 0.27, respectively)^19,40. During recurrent HK, patients were managed according to the severity of the episode, changes in the RAASi dose with either intermittent administration of SPS for three days or received SPS until reinitiation of SZC or patiromer. The probability of patients who experienced AEs under SPS was extracted from a study that compared calcium polystyrene sulfonate (CPS) and SPS to correct K + levels among CKD patients⁴¹.

RAASi use

Among the CKD cohort, the most common RAASis were ACEIs (ramipril or lisinopril) or ARBs (valsartan or telmisartan). The proportion of patients on ACEIs or ARBs was assumed to be 50% for each. The standardized RAASis for the HF cohort was MRA (spironolactone or eplerenone assumed 50% for each) administered either ACEI (ramipril or lisinopril) or ARB (valsartan).

With the initial administration of RAASis, a physical examination through serum K + and KFT as a baseline assessment should be performed. Initially, the patient starts on a suboptimal dose and increases the dose depending on the clinical response and tolerability. Any change in RAASis (up/downtitration or discontinuation) should be decided by specialists in secondary care once every 2 weeks. With each titration, the above tests should be conducted biweekly.

Sensitivity analysis

Deterministic and probabilistic sensitivity analyses were performed to assure the robustness of the results. Various parameters were varied with 10–20% above or below their base case values. The parameters tested were the utility data, clinical parameters and unit cost data for each treatment arm.

Results

Initial and recurrent HK at serum K± ≥ 5.1

Over a lifetime horizon, SZC showed a significant improvement in LYs and QALYs compared to patiromer. Within different scenarios of CKD/HF, SZC was a cost-saving option, with or without RRT, whether one-off administration or repeated administration, except for one-off treatment administration among the HF cohort, which generated an ICER of KWD 331/QALY (cost-effective; below 3xGPD per capita in Kuwait). The incremental QALY of SZC ranged from 0.007 to 0.202. In addition, the savings observed with SZC fall within a range of KWD −60 to KWD −1,235. Table 4 shows the base-case results at serum K+ ≥ 5.1.

Table 4. Model results at serum K+ ≥ 5.1.

Treatment Strategy	Total Cost (KWD)	Difference in Cost (KWD)	Total QALYs	Difference in QALYs	Total LYs	Difference in LYs	ICER	Interpretation
A. CKD cohort (without RRT), one-off administration of SZC/patiromer
SZC	KWD 11,210	KWD −60	4.634	0.009	7.036	0.014	Dominant	SZC generates a better QALY/LY when compared to patiromer in addition of being a cost saving
Patiromer	KWD 11,270		4.625		7.022
A. CKD cohort (without RRT), repeated administration of SZC/patiromer
SZC	KWD 13,857	KWD −1,235	4.769	0.097	7.246	0.161	Dominant	SZC generates a better QALY/LY when com-pared to patiromer in addition of being a cost saving
Patiromer	KWD 15,092		4.672		7.085
A. CKD cohort (with RRT), one-off administration of SZC/patiromer
SZC	KWD 13,219	KWD – 99	4.733	0.007	7.309	0.009	Dominant	SZC generates a better QALY/LY when com-pared to patiromer in addition of being a cost saving
Patiromer	KWD 13,318		4.726		7.300
A. CKD cohort (with RRT), repeated administration of SZC/patiromer
SZC	KWD 15,841	KWD −1,074	4.867	0.105	7.516	0.183	Dominant	SZC generates a better QALY/LY when com-pared to patiromer in adition of being a cost saving
Patiromer	KWD 16,915		4.762		7.333
A. HF cohort, repeated administration of SZC/patiromer
SZC	KWD 3,887	KWD −583	2.330	0.099	4.061	0.177	Dominant	SZC generates a better QALY/LY when com-pared to patiromer in addition of being a cost saving
Patiromer	KWD 4470		2.231		3.884
A. HF cohort, one-off administration of SZC/patiromer
SZC	KWD 2,537	KWD 67	2.038	0.202	3.546	0.357	KWD 331/QALY	Compared to patiromer, SZC is a cost-effective option
Patiromer	KWD 2,470		1.836		3.189

Abbreviations. KWD, Kuwaiti Dinar; QALYs, Quality Adjusted Life Years; ICER, Incremental Cost Effectiveness Ratio; LYs, life years; CKD, chronic kidney disease; HF, heart failure; RRT, renal replacement therapy; SZC, Sodium zirconium cyclosilicate.

Initial and recurrent HK at serum K± ≥ 5.5 mmol/L

For the HK scenario at serum K+ ≥ 5.5 mmol/L, SZC led to an improvement in the generated outcomes of QALYs and LYs over a lifetime horizon. Among different CKD/HF cohorts, SZC was shown to be a cost-saving option within all scenarios, with or without RRT, whether one-off administration or repeated administration. The incremental QALY of SZC ranged from 0.098 to 0.203, while the observed savings started from KWD −15 to KWD −1,448. Table S3 shows the base-case results at serum K+ ≥ 5.5.

Sensitivity analyses

The robustness of the model and the results were tested using deterministic sensitivity analyses (DSA). The tornado diagram (Figure 2) suggests that the results of the repeated administration of SZC versus patiromer in CKD patients at serum K+ ≥ 5.1 mmol/L was robust in one-way sensitivity analyses (or DSA). The most sensitive parameter in the model was the long-term post discontinuation management cost when we compared the repeated administration of SZC versus patiromer in CKD patients without RRT at serum K+ ≥ 5.1 mmol/L. Threshold analysis (Table 5) was performed on the most impactful parameters depicted in Figure 2, and therefore the decision was changed as the cost-effectiveness threshold for Kuwait is three x gross domestic product per capita (KWD 12,581 × 3= KWD 37,743).

Figure 2. One way sensitivity analysis of the repeated administration of SZC versus patiromer in CKD patients at serum K+ ≥ 5.1 mmol/L.

Table 5. Threshold analysis for the most influential parameters in the repeated administration of SZC versus patiromer in CKD patients at serum K+ ≥ 5.1 mmol/L.

Most influential Parameters	Base case value	Cut-off point	ICER	Interpretation
Long-term post-discontinuation management cost	KWD 15	KWD 1000	144,395	Conclusion changed^a
Baseline eGFR (ml/min/1.73m2)	47	51.7	−16,390	Conclusion did not change (cost saving)
CKD Annual eGFR decline - No RAASi	3.52	3.87	−14,160	Conclusion did not change (cost saving)
Health state utility: CKD 4	0.85	0.77	−13,491	Conclusion did not change (cost saving)

^a The cost-effectiveness threshold for Kuwait (three × GDP per capita; KWD 37,743).

A cost-effectiveness plane was performed to measure the uncertainty around the parameters. All data inputs were simultaneously varied within different distributions. Figure 3 illustrates the 1,000-iteration cost outcome difference pairs. As shown in HF patients, most difference pairs are found in the northeast and southeast quadrants of the cost effectiveness plane, which indicates the value of one-off treatment administration of SZC use in HF patients. The cost-effectiveness acceptability curve over a range of willingness to pay thresholds is drawn in Figure 4.

Figure 3. Cost-effectiveness plane of the one-off administration of SZC versus patiromer in HF patients at serum K+ ≥ 5.1 mmol/L.

Figure 4. Cost-effectiveness acceptability curve of the one-off administration of SZC versus patiromer in HF patients at serum K+ ≥ 5.1 mmol/L.

Discussion

Our model simulated the economic and clinical outcomes of SZC against Patiromer for the correction of acute HK and maintenance of serum K + levels among patients with CKD/HF. This study simulated the natural history of CKD/HF progression, expected fluctuations in serum K + and modifications of RAASi doses. We aimed to inform policy-makers about the value of SZC in Kuwait to assure resource allocations and the avoidance of financial hardship.

Currently, the intermittent administration of SPS/CPS with downtitration or discontinuation of RAASis is used to manage HK, but these have significant limitations. There are limited long-term safety and efficacy studies for SPS/CPS^41,42. Serious gastrointestinal (GI) AEs and poor palatability have been linked to SPS^43,44. As a result, the medium- or long-term administration of SPS is not recommended^44,45. Unlike SPS/CPS, SZC showed good tolerability over a follow-up period of 12 months, which supports its long-term administration to manage HK¹⁹.

Different systematic reviews and meta-analyses of novel K + binders, such as SZC and patiromer, reflected that SZC is the drug of choice to manage patients with acute HK due to its rapid reduction in serum K+^46,47. After 48 h on SZC, the change of serum K + level was −0.67 mEq/L, while the reduction of serum K + level after 72 h of patiromer was −0.36 mEq/L⁴⁷. The rapid correction of serum K + with SZC during acute HK helps patients avoid HK complications such as arrhythmia^1,9.

SZC demonstrated significant efficacy and favourable safety in maintaining serum K + levels for long-term management⁴⁸. During the maintenance phase, the reduction in the serum K + level for SZC on Day 29 was −0.75 mmol/L⁴⁸, while it was −0.70 mEq/L at 4 weeks with a patiromer⁴⁷. A recent meta-analysis demonstrated that some AE rates of SZC were similar to those of placebo⁴⁷. Compared with SZC, patients had frequent GI AEs and hypomagnesemia⁴⁷.

Downtitration or discontinuation of RAASis prevents CKD/HF patients benefit from RAASis, and these patients experience an increased risk of adverse outcomes²⁸. SZC helps CKD/HF patients with continuation or optimization of RAASi therapy¹⁹. The reduction in serum K + levels with SZC does not require any changes in RAASi doses¹⁹. Spinowitz et al. demonstrated that among those administered SZC, 87% maintained or increased the RAASi doses¹⁹. Therefore, SZC allows patients to obtain clinical benefits from RAASis, which subsequently improve patients’ QoL⁴⁹.

There are some benefits of SZC to health care providers and patients who should be considered. Patiromer requires special precautions for transportation and storage (refrigerated at 2 °C: 8 °C)⁵⁰. In contrast, SZCs can be stored at room temperature without the need for any special storage conditions⁵¹, which ease dispensing and logistics services. In addition, SZC provides flexible dosing. If the patient missed an SZC dose, the next dose can be taken at the normal time⁵¹. Unlike SZC, the missed dose with the patiromer device should be taken as soon as possible on the same day⁵⁰. Moreover, SZC can be administered for patients with chronic haemodialysis⁵¹, while patiromer is not indicated for patients on dialysis.

A crucial consideration with patiromer is potential drug–drug interactions, which can be avoided by the administration of other oral medications, such as ACEIs/ARBs, at least 3 h before or after patiromer⁵². Therefore, adherence to other medications might be an obstacle for some patients⁴⁶.

In addition to promising outcomes with long-term administration of SZC, it can also reduce the total costs of HK^48,53. A study compared the health care resource utilization (HRU) associated with short-term (≤90 d) versus long-term (>90 d) SZC therapy⁵³. Long-term SZC therapy led to reductions of 33.0% and 23.3% for HK-related and all-cause hospitalization, respectively⁵³.

A cost-effectiveness analysis conducted in Norway and Sweden evaluated SZC against usual care (intermittent SPS/CPS therapy with RAASi dose adjustments) to treat HK among patients with CKD over a lifetime horizon⁵⁴. The study concluded that SZC was a cost-effective option for the management of HK, with an ICER of €14,838/QALY in Norway and €14,352/QALY in Sweden, as SZC led to a reduction in HK events and hospitalization⁵⁴.

This study can play an important role in the guidance of decision-makers and clinicians to choose the optimal regimen for the management of HK and maintain normal serum K + among CKD/HF patients. Our study has several strengths that must be mentioned. First, we considered SZC against another novel oral K + binder, patiromer. In addition, the K + trajectories of SZC and patiromer were extracted from strong clinical trials of the HARMONIZE trial (ZS-004) and OPAL-HK. Furthermore, we considered the natural progression of CKD/HF over a lifetime horizon. The model simulated different scenarios of acute and recurrent HK if the serum K + level exceeded either ≥ 5.1 or ≥5.5 mmol/L due to a lack of consent within experts about the appropriate K + threshold to initiate treatment. There are some limitations to this study that need to be discussed. The clinical trials of patiromer and SZC had broad exclusion criteria, and indirect costs were not considered. Calculation of indirect costs would not change the conclusion of the study since SZC was shown to reduce the complications and hospitalization that consequently affect the productivity loss so it leads to the same conclusion direction. The current results could be generalizable only to Kuwaiti public hospitals.

Conclusion

The evidence generated by our model recommends the inclusion of SZC as a treatment option to correct HK (at a serum K + level 5.5 mmol/L or ≥ 5.1 mmol/L) and maintain normal serum K + level for CKD/HF patients within Kuwaiti healthcare system. The drug acquisition cost of SZC offsets the cost saved from reducing frequent HK episodes, RAASis discontinuation/down-titration, MACE and hospitalization.

Transparency

• • •

Author contributions

GE: conducting the analysis and writing the manuscript. MI: Collection and interpretation of data, and revision of manuscript. KA and AA: retrieving data, local clinical practice validation and writing the manuscript. All authors contributed to the article and approved the submitted version.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgements

The authors gratefully acknowledge Mariam Elattar for the writing assistance utilized in the production of this manuscript.

Declaration of financial/other relationships

GE was employed by HTA Office, LLC. GE was a speaker in different pharmaceutical companies. The authors declare that the fund received was for the submission and the open access publication fees. The authors have no other financial relationships to disclose.

Additional information

Funding

This study was funded by AstraZeneca, Emirates, with no involvement in the study design, analysis, interpretation of results or manuscript writing.