Abstract
Rationale & Objective
Matched real-word evidence studies characterizing the relationship between duration of sodium zirconium cyclosilicate (SZC) therapy and hyperkalemia (HK)-related health care resource utilization (HRU) outcomes have been limited. In this GALVANIZE Outcome matched real-word evidence study, we compared HRU between matched short-term and long-term SZC users overall and in key sub-studies.
Study Design
A retrospective, matched, observational cohort study.
Setting & Population
Adults initiating outpatient SZC identified from a large United States claims database (July 2018-December 2022).
Exposures
Short-term SZC use (≤30 days) versus long-term SZC use (>90 days).
Outcomes
The composite primary outcome was HK-related hospitalizations or emergency department (ED) visits. Other outcomes included HK-related and all-cause hospitalizations, ED visits, and outpatient visits.
Analytical Approach
Short-term and long-term SZC users were exactly and propensity score matched. Outcomes were compared during follow-up using generalized estimating equations. Sub-studies were conducted among renin–angiotensin–aldosterone system inhibitor users, patients with renal conditions, Medicare Advantage beneficiaries, and Medicaid beneficiaries.
Results
Among 3,133 eligible matched pairs (mean age 64 years; 58% male; 91% with kidney disease), patients with long-term versus short-term SZC use had a 40% lower rate of HK-related hospitalizations or ED visits (incidence rate ratio [IRR] 0.60 [95% CI, 0.52-0.70]), a 36% lower rate of HK-related hospitalizations (IRR 0.64 [95% CI, 0.55-0.74]), and a 49% lower rate of HK-related ED visits (IRR 0.51 [95% CI, 0.37-0.70]) during follow-up (all P < 0.001). The rates of HK-related outpatient visits were similar between short-term and long-term SZC users (3.22 vs 2.72; IRR 0.85 [95% CI, 0.71-1.01]; P = 0.059). Similar trends were observed for all-cause HRU and across sub-studies.
Limitations
Claims data contain limited clinical information; residual confounding factors remain possible.
Conclusions
Patients with long-term SZC use had significantly lower rates of HK-related hospitalizations or ED visits than matched patients with short-term SZC use.
Index Words: Duration of therapy, health care resource utilization, hospitalization, hyperkalemia, sodium zirconium cyclosilicate
Plain-Language Summary
This study looked at whether people who stayed on the potassium-lowering medication sodium zirconium cyclosilicate (SZC) for a longer time had fewer hospital or emergency room visits for high potassium (hyperkalemia). Using United States health insurance data from 2018-2022, researchers compared over 3,000 matched pairs of adults who used SZC for a short time (30 days or less) or a long time (more than 90 days). People who took SZC long-term had about 40% fewer hospital or emergency room visits related to high potassium than those who used SZC short-term. These results were consistent across patient subgroups. Overall, this study suggests that continuing SZC may be associated with a lower rate of hospital visits for people with hyperkalemia.
Hyperkalemia (HK), defined as serum potassium > 5.0 nmol/L, is an established complication in patients with reduced kidney function1,2 that is associated with increased risk of kidney events, life-threatening cardiac dysrhythmia, and overall mortality.3, 4, 5, 6 In the United States (US), the estimated prevalence of HK is ∼1.6% and is higher among individuals with chronic kidney disease (CKD), heart failure (HF), diabetes, and hypertension.1,7 Hyperkalemia may arise from high dietary potassium intake, acidosis, or altered kidney potassium handling.1,8 In addition, the use of renin–angiotensin–aldosterone system inhibitors (RAASis; eg, angiotensin-converting enzyme inhibitors, angiotensin-receptor blockers, and mineralocorticoid receptor antagonists), a cornerstone treatment of various kidney and cardiovascular diseases,8,9 has been shown to increase the frequency of HK.10,11
Treatment options for HK include potassium-binding treatments, such as sodium polystyrene sulfonate, patiromer, and sodium zirconium cyclosilicate (SZC), as well as temporizing agents such as sodium bicarbonate, calcium gluconate, calcium dextrose, and dextrose/insulin.9,12 Specifically, SZC is a selective potassium binder approved in the US in May 2018 for the treatment of adults with HK.13
Previous studies have demonstrated increased health care resource utilization (HRU) and cost burden among patients with HK relative to their counterparts without HK, including patients with kidney disease.14, 15, 16, 17, 18 Furthermore, the duration of HK treatments has also been suggested to influence HRU. In the RECOGNIZE I real-world evidence (RWE) study, patients with long-term outpatient SZC therapy had a 33% lower proportion of HK-related hospitalizations compared with patients with short-term outpatient SZC therapy.19 However, the RECOGNIZE I study was descriptive and did not adjust for differences in patient characteristics between long-term and short-term SZC users.
To date, there have been few matched RWE studies characterizing the relationship between duration of SZC therapy and HRU outcomes. In addition, little research exists that quantifies HK-related HRU among patients with concomitant RAASi use. In this GALVANIZE Outcome RWE study, we assessed HK-related and all-cause HRU associated with short-term versus long-term outpatient SZC use in matched cohorts of SZC users. We conducted sub-studies among patients with RAASi use at the time of SZC initiation, patients with renal complications, patients with Medicare Advantage coverage, and patients with Medicaid coverage.
Methods
Study Design
The GALVANIZE Outcome RWE study was a noninterventional, retrospective, matched, observational cohort study of patients treated with outpatient SZC therapy. Patients with short-term SZC use (≤30 days) were compared with matched patients with long-term SZC use (>90 days). Definitions of short-term versus long-term use were based on clinical input, typical SZC prescription quantities, and prior literature.
The study used data from a large, closed US claims database that were fully adjudicated by payers from 2 linked data sources, provided by HealthVerity, spanning July 2018-December 2022. The database included medical and pharmacy data from more than 150 unique payers and 115 million patients.
Study Population
The overall study population included patients aged ≥18 years with a record for SZC in the outpatient setting from closed pharmacy claims (the index date was defined as the date of the first record occurring from January 1, 2019, to December 1, 2022). Patients were required to have continuous medical and pharmacy enrollment for ≥6 months before the index date (baseline period) and ≥30 days after the index date.
Patients were followed for a maximum of 6 months after the index date (follow-up period). Based on the sum of days of supply of SZC before discontinuation (defined as ≥30-day gap without a new SZC fill after the end of the prior fill) during the follow-up period, patients were classified into the short-term SZC use cohort (≤30 days) or the long-term SZC use cohort (>90 days). Patients were censored at the reinitiation of SZC after discontinuation (to minimize the risk of reverse causation, ie, certain HRU may prolong SZC use), at the initiation of another potassium binder, at the end of data availability, death, or 6 months after the index date, whichever occurred first.
Matching
Patients in the short-term SZC use cohort were matched 1:1 to patients in the long-term SZC use cohort using a 2-step process. First, we exactly matched patients on the following baseline characteristics: kidney failure (diagnosis code for kidney failure or diagnosis code for CKD stage 5/kidney failure with receipt of hemodialysis), CKD (any stage), HF, HK diagnosis code, acute kidney injury, any RAASi use on the index date, Medicare Advantage coverage, Medicaid coverage, and any inpatient admission during baseline. Second, we used a logistic regression model predicting short-term versus long-term SZC use to generate the propensity score. The final model included the following variables: demographics (age, sex, region, and index year), any baseline RAASi use, patiromer use, sodium polystyrene sulfonate use, ranked CKD stage, hypertension, COVID-19, Charlson Comorbidity Index (continuous), all variables used for exact matching, and the top 20 ranked covariates from the propensity score ranking algorithm not included above. We then matched patients 1:1 within each coarsened exact match category using the propensity score and caliper matching with a caliper of 0.05.
Sub-studies
A sub-study was conducted among patients with RAASi use overlapping with the index date. Supplemental sub-studies were also conducted separately among patients with CKD or kidney failure, patients with kidney failure, patients with a HK diagnosis, patients with Medicare Advantage coverage, and patients with Medicaid coverage.
Variables and Outcomes
Patient characteristics including demographics, comorbid conditions, treatments, other clinical characteristics, and HK-related and all-cause HRU were assessed during baseline for the short-term SZC use and long-term SZC use cohorts after matching.
The composite primary outcome was HK-related hospitalizations or emergency department (ED) visits during the follow-up period. Hospitalizations or ED visits with a diagnosis of HK in any position were considered HK-related. The secondary outcome was HK-related hospitalizations during follow-up. Exploratory outcomes included composite all-cause hospitalizations or ED visits, all-cause hospitalizations, HK-related and all-cause ED visits, and HK-related and all-cause outpatient visits during follow-up.
Statistical Analyses
Patient baseline characteristics and HRU outcomes were summarized descriptively using means and standard deviations (SDs) for continuous variables; and counts and percentages for categorical variables. Matched baseline characteristics were compared between the short-term and long-term SZC use cohorts using standardized mean differences (SMDs; <0.20 was considered well-matched).20 The HRU outcomes were compared between the cohorts using generalized estimating equations to account for matching. Outcomes were reported as rates per person-year (PPY) and incidence rate ratios (IRR) with corresponding 95% CIs.
For the RAASi at index sub-study, median time to RAASi discontinuation and the proportion of patients with RAASi discontinuation at different time points were summarized and compared between patients in the matched short-term SZC use and long-term SZC use cohorts. RAASi discontinuation was defined as ≥90-day gap without a new RAASi fill after the end of the prior fill. The discontinuation date was the last date covered by days of supply for the RAASi record before discontinuation. Patients were censored if they did not have ≥90 days of continuous enrollment following the treatment end date, as patients may have continued or discontinued treatment outside of the data. Kaplan-Meier analyses were used to obtain median times to RAASi discontinuation, and adjusted Cox proportional hazards models were used to compare time to RAASi discontinuation between the matched cohorts.
All analyses were conducted using R 3.6.3. A P-value < 0.05 was considered statistically significant.
Ethics
The study was considered exempt research under 45 CFR § 46.104(d)(4), and informed consent was not required as the study involved only the secondary use of data that were de-identified in adherence with the Health Insurance Portability and Accountability Act, specifically, 45 CFR § 164.514. Patient data were used in accordance with the Helsinki Declaration as revised in 2013.
Results
Sample Selection
A total of 18,454 patients met the study criteria before matching, including 9,629 short-term SZC users and 4,452 long-term SZC users (Fig 1). After matching, 6,266 patients were included in the overall sample, with 3,133 matched pairs in each of the short-term SZC use and long-term SZC use cohorts. For the RAASi at index sub-study, 1,586 matched pairs were included in each cohort. For the supplemental sub-studies, the number of matched pairs included in each cohort was 2,854 for CKD or kidney failure; 1,010 for kidney failure; 2,088 for HK diagnosis; 1,593 for Medicare Advantage coverage; and 647 for Medicaid coverage.
Figure 1.
Sample selection. CKD, chronic kidney disease; HK, hyperkalemia; RAASi, renin–angiotensin–aldosterone system inhibitor; SZC, sodium zirconium cyclosilicate.
Baseline Characteristics
In the overall sample, baseline patient characteristics were similar between the short-term SZC use and long-term SZC use cohorts after matching (Table 1). In both cohorts, the mean age was 64 years, and there were slightly more male patients (59% and 58%, respectively) (both SMD < 0.02). The mean Charlson Comorbidity Index was 3.1 in the short-term SZC use cohort and 3.2 in the long-term SZC use cohort (SMD = 0.02). In both cohorts, 91% of patients had CKD or kidney failure, 32% had kidney failure, and 67% had HK during baseline. Other common comorbid conditions included hypertension (short-term: 86% vs long-term: 86%), any diabetes (67% vs 70%), HF (both 30%), and acute kidney injury (both 29%) (all SMD < 0.05).
Table 1.
Baseline Characteristics in the Overall Matched Sample
| Overall Matched Sample |
RAASi at Index Sub-study |
|||||
|---|---|---|---|---|---|---|
| Short-Term SZC |
Long-Term SZC |
SMD | Short-Term SZC |
Long-Term SZC |
SMD | |
| n = 3,133 | n = 3,133 | n = 1,586 | n = 1,586 | |||
| Demographics | ||||||
| Age on index date (y), mean ± SD | 64.1 ± 15.9 | 64.2 ± 15.3 | 0.011 | 65.3 ± 15.1 | 65.7 ± 14.8 | 0.024 |
| Female sex, n (%) | 1,301 (41.5%) | 1,319 (42.1%) | 0.012 | 644 (40.6%) | 655 (41.3%) | 0.014 |
| Region, n (%) | 0.020 | 0.042 | ||||
| Northeast | 895 (28.6%) | 890 (28.4%) | 449 (28.3%) | 454 (28.6%) | ||
| Midwest | 526 (16.8%) | 548 (17.5%) | 256 (16.1%) | 249 (15.7%) | ||
| South | 980 (31.3%) | 967 (30.9%) | 493 (31.1%) | 488 (30.8%) | ||
| West | 702 (22.4%) | 696 (22.2%) | 372 (23.5%) | 372 (23.5%) | ||
| Other/Unknown | 30 (1.0%) | 32 (1.0%) | 16 (1.0%) | 23 (1.5%) | ||
| Insurance type (medical), n (%) | 0.000 | 0.013 | ||||
| Commercial | 856 (27.3%) | 856 (27.3%) | 419 (26.4%) | 421 (26.5%) | ||
| Medicaid | 647 (20.7%) | 647 (20.7%) | 306 (19.3%) | 306 (19.3%) | ||
| Medicare advantage | 1,593 (50.8%) | 1,593 (50.8%) | 844 (53.2%) | 844 (53.2%) | ||
| Unknown | 37 (1.2%) | 37 (1.2%) | 17 (1.1%) | 15 (0.9%) | ||
| Index year, n (%) | 0.028 | 0.086 | ||||
| 2019 | 243 (7.8%) | 240 (7.7%) | 129 (8.1%) | 135 (8.5%) | ||
| 2020 | 605 (19.3%) | 624 (19.9%) | 289 (18.2%) | 317 (20.0%) | ||
| 2021 | 1,094 (34.9%) | 1,055 (33.7%) | 570 (35.9%) | 507 (32.0%) | ||
| 2022 | 1,191 (38.0%) | 1,214 (38.7%) | 598 (37.7%) | 627 (39.5%) | ||
| CCI, mean ± SD | 3.1 ± 2.2 | 3.2 ± 2.1 | 0.021 | 3.0 ± 2.0 | 3.1 ± 2.0 | 0.011 |
| CKD, n (%) | ||||||
| Any CKD or kidney failure | 2,854 (91.1%) | 2,854 (91.1%) | 0.000 | 1,458 (91.9%) | 1,458 (91.9%) | 0.000 |
| Any CKD | 1,844 (58.9%) | 1,844 (58.9%) | 0.000 | 1,136 (71.6%) | 1,136 (71.6%) | 0.000 |
| Kidney failurea | 1,010 (32.2%) | 1,010 (32.2%) | 0.000 | 322 (20.3%) | 322 (20.3%) | 0.000 |
| Other comorbid conditions | ||||||
| Acute kidney injury, n (%) | 909 (29.0%) | 909 (29.0%) | 0.000 | 397 (25.0%) | 397 (25.0%) | 0.000 |
| HF, n (%) | 942 (30.1%) | 942 (30.1%) | 0.000 | 476 (30.0%) | 476 (30.0%) | 0.000 |
| HK diagnosis code, n (%) | 2,088 (66.6%) | 2,088 (66.6%) | 0.000 | 1,021 (64.4%) | 1,021 (64.4%) | 0.000 |
| Number of HK diagnoses, mean ± SD | 4.4 ± 11.2 | 4.3 ± 9.7 | 0.007 | 2.9 ± 7.5 | 3.2 ± 8.1 | 0.038 |
| COVID-19 diagnosis, n (%) | 298 (9.5%) | 296 (9.4%) | 0.002 | 125 (7.9%) | 111 (7.0%) | 0.034 |
| Any diabetes, n (%) | 2,110 (67.3%) | 2,177 (69.5%) | 0.046 | 1,138 (71.8%) | 1,190 (75.0%) | 0.074 |
| Type I diabetes | 265 (8.5%) | 276 (8.8%) | 0.013 | 142 (9.0%) | 137 (8.6%) | 0.011 |
| Type II diabetes | 2,089 (66.7%) | 2,158 (68.9%) | 0.047 | 1,125 (70.9%) | 1,178 (74.3%) | 0.075 |
| Hypertension, n (%) | 2,698 (86.1%) | 2,706 (86.4%) | 0.007 | 1,434 (90.4%) | 1,446 (91.2%) | 0.026 |
| Resistant hypertensionb, n (%) | 732 (23.4%) | 776 (24.8%) | 0.033 | 535 (33.7%) | 569 (35.9%) | 0.045 |
| HK treatments, n (%) | ||||||
| Patiromer | 178 (5.7%) | 177 (5.6%) | 0.001 | 80 (5.0%) | 71 (4.5%) | 0.027 |
| SPS | 296 (9.4%) | 343 (10.9%) | 0.050 | 137 (8.6%) | 171 (10.8%) | 0.072 |
| Dialysis | 87 (2.8%) | 79 (2.5%) | 0.016 | 29 (1.8%) | 26 (1.6%) | 0.014 |
| Any temporizing agent | 119 (3.8%) | 159 (5.1%) | 0.062 | 54 (3.4%) | 75 (4.7%) | 0.067 |
| Calciumc (IV) | 3 (0.1%) | 6 (0.2%) | 0.025 | 1 (0.1%) | 2 (0.1%) | 0.021 |
| Glucose + insulin (IV) | 7 (0.2%) | 10 (0.3%) | 0.018 | 3 (0.2%) | 4 (0.3%) | 0.013 |
| Sodium bicarbonate (IV or oral) | 112 (3.6%) | 149 (4.8%) | 0.059 | 50 (3.2%) | 71 (4.5%) | 0.069 |
| Any diuretic | 1,497 (47.8%) | 1,526 (48.7%) | 0.019 | 902 (56.9%) | 898 (56.6%) | 0.005 |
| Loop diuretics | 1,104 (35.2%) | 1,173 (37.4%) | 0.046 | 584 (36.8%) | 622 (39.2%) | 0.049 |
| Thiazides and thiazide-like diuretics | 482 (15.4%) | 441 (14.1%) | 0.037 | 356 (22.4%) | 320 (20.2%) | 0.055 |
| Potassium-sparing diuretics | 264 (8.4%) | 278 (8.9%) | 0.016 | 228 (14.4%) | 221 (13.9%) | 0.013 |
| RAASi use, n (%) | ||||||
| Any RAASi during baseline | 1,841 (58.8%) | 1,832 (58.5%) | 0.006 | 1,572 (99.1%) | 1,554 (98.0%) | 0.095 |
| ACEi | 954 (30.5%) | 892 (28.5%) | 0.043 | 816 (51.5%) | 777 (49.0%) | 0.049 |
| ARB | 881 (28.1%) | 930 (29.7%) | 0.035 | 765 (48.2%) | 803 (50.6%) | 0.048 |
| DRI | 3 (0.1%) | 1 (0.0%) | 0.025 | 2 (0.1%) | 1 (0.1%) | 0.021 |
| ARNI | 60 (1.9%) | 101 (3.2%) | 0.083 | 54 (3.4%) | 89 (5.6%) | 0.107 |
| MRA | 248 (7.9%) | 264 (8.4%) | 0.019 | 218 (13.7%) | 211 (13.3%) | 0.013 |
| Any RAASi at index | 1,586 (50.6%) | 1,586 (50.6%) | 0.000 | 1,586 (100.0%) | 1,586 (100.0%) | 0.000 |
| ACEi | 929 (29.7%) | 887 (28.3%) | 0.030 | 929 (58.6%) | 887 (55.9%) | 0.054 |
| ARB | 868 (27.7%) | 918 (29.3%) | 0.035 | 868 (54.7%) | 918 (57.9%) | 0.064 |
| DRI | 2 (0.1%) | 1 (0.0%) | 0.015 | 2 (0.1%) | 1 (0.1%) | 0.021 |
| ARNI | 78 (2.5%) | 111 (3.5%) | 0.062 | 78 (4.9%) | 111 (7.0%) | 0.088 |
| MRA | 277 (8.8%) | 266 (8.5%) | 0.012 | 277 (17.5%) | 266 (16.8%) | 0.018 |
| Other medication use, n (%) | ||||||
| β Blockers | 1,797 (57.4%) | 1,889 (60.3%) | 0.060 | 948 (59.8%) | 984 (62.0%) | 0.047 |
| SGLT2i monotherapy | 193 (6.2%) | 243 (7.8%) | 0.063 | 157 (9.9%) | 193 (12.2%) | 0.072 |
| Combination therapy including SGLT2i | 7 (0.2%) | 10 (0.3%) | 0.018 | 7 (0.4%) | 9 (0.6%) | 0.018 |
| Calcium channel blockers | 1,662 (53.0%) | 1,654 (52.8%) | 0.005 | 880 (55.5%) | 864 (54.5%) | 0.020 |
| HK-related HRU, n (%) | ||||||
| Any HK-related hospitalization or ED | 1,047 (33.4%) | 942 (30.1%) | 0.072 | 422 (26.6%) | 367 (23.1%) | 0.080 |
| Any HK-related hospitalization | 875 (27.9%) | 833 (26.6%) | 0.030 | 332 (20.9%) | 316 (19.9%) | 0.025 |
| Any HK-related ED | 356 (11.4%) | 280 (8.9%) | 0.080 | 158 (10.0%) | 112 (7.1%) | 0.104 |
| Any HK-related outpatient | 1,192 (38.0%) | 1,348 (43.0%) | 0.102 | 657 (41.4%) | 717 (45.2%) | 0.076 |
| All-cause HRU, n (%) | ||||||
| Any all-cause hospitalization or ED | 1,648 (52.6%) | 1,601 (51.1%) | 0.030 | 740 (46.7%) | 704 (44.4%) | 0.046 |
| Any all-cause hospitalization | 1,178 (37.6%) | 1,178 (37.6%) | 0.000 | 469 (29.6%) | 469 (29.6%) | 0.000 |
| Any all-cause ED | 1,100 (35.1%) | 1,022 (32.6%) | 0.053 | 528 (33.3%) | 464 (29.3%) | 0.087 |
| Any all-cause outpatient | 2,905 (92.7%) | 2,909 (92.9%) | 0.005 | 1,506 (95.0%) | 1,491 (94.0%) | 0.041 |
Note: SMDs < 0.2 indicate well-matched cohorts.
Abbreviations: ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin-receptor blocker; ARNI, angiotensin-receptor neprilysin inhibitor; CCI, Charlson Comorbidity Index; CKD, chronic kidney disease; DRI, direct renin inhibitor; ED, emergency department; HF, heart failure; HK, hyperkalemia; HRU, health care resource utilization; IV, intravenous therapy; MRA, mineralocorticoid receptor antagonist; RAASi, renin–angiotensin–aldosterone system inhibitor; SD, standard deviation; SGLT2i, sodium-glucose cotransporter-2 inhibitors; SMD, standardized mean difference; SPS, sodium polystyrene sulfonate; SZC, sodium zirconium cyclosilicate.
Kidney failure was defined by either any diagnosis code for kidney failure or a diagnosis code for CKD stage 5 or kidney failure (unspecified) only (ie, no additional CKD stage 5 or kidney failure-specific diagnosis codes).
Resistant hypertension was defined as the presence of a hypertension diagnosis code and the use of 3 or more classes of anti-hypertensive medications, including a diuretic. The hypertension drug classes included β blockers, calcium channel blockers, α blockers, central α-2 receptor agonists, vasodilators, ACEi, ARB, ARNI, DRI, MRA, and diuretics (including loop diuretics, potassium-sparing diuretics, and thiazides and thiazide-like diuretics).
Calcium includes calcium gluconate and calcium dextrose.
Among both short-term and long-term SZC users, half of the patients (51%) had RAASi use on the index date, and the most used RAASis were angiotensin-converting enzyme inhibitor (short-term: 30% vs long-term: 28%) and angiotensin-receptor blocker (28% vs 29%) (both SMD < 0.05). Other treatment use was similar between the 2 cohorts, including β blockers (57% vs 60%) and calcium channel blockers (53% vs 53%) (both SMD < 0.1).
Baseline HK-related and all-cause HRU was generally well-balanced between short-term and long-term SZC users (SMD ≤ 0.1). Overall, nearly one-third (32%) had an HK-related hospitalization or ED visit during baseline, and approximately half of patients (52%) had any all-cause hospitalization or ED visit during baseline (both SMD < 0.1).
Baseline characteristics of matched patients in the RAASi at the index sub-study (Table 1) and in the respective supplemental sub-studies (Tables S1-S5) were similar to those of the overall matched sample.
HRU Outcomes
Overall Sample
In the follow-up period, which had a maximum duration of 6 months, patients with short-term SZC use had a mean ± SD (median [25th-75th percentile]) follow-up of 135.0 ± 57.3 (180.0 [83.0-182.6]) days and SZC therapy duration of 20.7 ± 11.6 (30.0 [9.0-30.0]) days, whereas patients with long-term SZC use had a mean ± SD (median [25th-75th percentile]) follow-up of 171.8 ± 23.7 (182.6 [182.6-182.6]) days and SZC therapy duration of 167.7 ± 52.8 (167.0 [128.0-189.0]) days (both P < 0.001).
During follow-up, compared with patients with short-term SZC use, those with long-term SZC use had a 40% lower rate of HK-related hospitalizations or ED visits (short-term: 1.20 PPY vs long-term: 0.72 PPY; IRR 0.60 [95% CI, 0.52-0.70]), a 36% lower rate of HK-related hospitalizations (0.87 vs 0.55; IRR 0.64 [95% CI, 0.55-0.74]), and a 49% lower rate of HK-related ED visits (0.33 vs 0.17; IRR 0.51 [95% CI, 0.37-0.70]) (all P < 0.001; Fig 2A). The rates of HK-related outpatient visits were similar between short-term and long-term SZC users (3.22 vs 2.72; IRR 0.85 [95% CI, 0.71-1.01]; P = 0.06).
Figure 2.
Comparison of HRU rates during follow-up by duration of SZC use in the overall sample. CI, confidence interval; ED, emergency department; HK, hyperkalemia; HRU, health care resource utilization; IRR, incidence rate ratio; SD, standard deviation; SZC, sodium zirconium cyclosilicate. aThe mean ± SD duration of follow-up was 135.0 ± 57.3 days among short-term SZC users and 171.8 ± 23.7 days among long-term SZC users (P < 0.001). bThe mean ± SD SZC therapy duration was 20.7 ± 11.6 days for short-term SZC users and 167.7 ± 52.8 days for long-term SZC users (P < 0.001).
The rates of all-cause HRU during follow-up exhibited similar trends. Relative to patients with short-term SZC use, patients with long-term SZC use had a 31% lower rate of all-cause hospitalizations or ED visits (short-term: 3.37 PPY vs long-term: 2.33 PPY; IRR 0.69 [95% CI, 0.63-0.75]), a 33% lower rate of all-cause hospitalizations (1.84 vs 1.23; IRR 0.67 [95% CI, 0.60-0.74]), and a 28% lower rate of all-cause ED visits (1.53 vs 1.10; IRR 0.72 [95% CI, 0.64-0.81]) (all P < 0.001; Fig 2B). The rates of all-cause outpatient visits were similar between short-term and long-term SZC users (26.81 vs 26.45; IRR 0.99 [95% CI, 0.93-1.04]; P = 0.64).
Sub-study Among Patients with RAASi Use at Index
In the RAASi at index sub-study, patients with short-term SZC use had a shorter mean ± SD follow-up of 136.6 ± 56.1 days than that for patients with long-term SZC use at 172.1 ± 23.8 days (P < 0.001). The mean ± SD SZC therapy duration was 20.1 ± 11.8 days among short-term SZC users and 170.6 ± 52.8 days among long-term SZC users (P < 0.001).
The rates of HK-related HRU during follow-up were significantly lower among patients with long-term versus short-term SZC use in all categories in the RAASi at index sub-study (Fig 3A). Compared with short-term SZC users, long-term SZC users had a 44% lower rate of HK-related hospitalizations or ED visits (short-term: 0.98 PPY vs long-term: 0.55 PPY; IRR 0.56 [95% CI, 0.44-0.70]), a 41% lower rate of HK-related hospitalizations (0.69 vs 0.40; IRR 0.59 [95% CI, 0.48-0.73]), a 52% lower rate of HK-related ED visits (0.30 vs 0.14; IRR 0.48 [95% CI, 0.29-0.80]), and a 22% lower rate of HK-related outpatient visits (3.47 vs 2.71; IRR 0.78 [95% CI, 0.62-0.99]) (all P < 0.05).
Figure 3.
Comparison of HRU rates during follow-up by duration of SZC use in the RAASi at index sub-study. CI, confidence interval; ED, emergency department; HK, hyperkalemia; HRU, health care resource utilization; IRR, incidence rate ratio; RAASi, renin–angiotensin–aldosterone system inhibitor; SD, standard deviation; SZC, sodium zirconium cyclosilicate. aThe mean ± SD duration of follow-up was 136.6 ± 56.1 days among short-term SZC users and 172.1 ± 23.8 days among long-term SZC users (P < 0.001). bThe mean ± SD SZC therapy duration was 20.1 ± 11.8 days for short-term SZC users and 170.6 ± 52.8 days for long-term SZC users (P < 0.001).
For all-cause HRU (Fig 3B), long-term SZC users had a 33% lower rate of all-cause hospitalizations or ED visits than short-term SZC users (short-term: 3.06 PPY vs long-term: 2.06 PPY; IRR 0.67 [95% CI, 0.59-0.77]), a 36% lower rate of all-cause hospitalizations (1.63 vs 1.04; IRR 0.64 [95% CI, 0.54-0.75]), and a 28% lower rate of all-cause ED visits (1.43 vs 1.02; IRR 0.72 [95% CI, 0.59-0.86]) (all P < 0.001). The rates all-cause outpatient visits were similar between short-term and long-term SZC users (26.18 vs 25.87; IRR 0.99 [95% CI, 0.92-1.06]; P = 0.75).
Patients with long-term SZC use had a 34% lower rate of RAASi discontinuation than those with short-term SZC use (hazard ratio 0.66 [95% CI, 0.57-0.76], P < 0.001; Fig 4). The median time to RAASi discontinuation was not reached in either cohort.
Figure 4.
Time to RAASi discontinuation in the RAASi at index sub-study. CI, confidence interval; HR, hazard ratio; RAASi, renin–angiotensin–aldosterone system inhibitor.
Discussion
In this retrospective matched cohort study of patients receiving outpatient SZC therapy, patients with long-term SZC use had a 36%-49% lower rate of HK-related hospitalizations and/or ED visits and 28%-33% lower rates of all-cause hospitalizations or ED visits than patients with short-term SZC use over the up to 6-month follow-up. The results were consistent across patients with RAASi use at SZC initiation, patients with CKD or kidney failure, kidney failure, and HK diagnoses, as well as patients with Medicare Advantage or Medicaid coverage. These results suggest that long-term SZC therapy may be associated with a lower risk of acute HK-related health events that necessitate hospitalizations or ED visits when compared with short-term SZC therapy.
The findings of this study are consistent with the descriptive results from the RECOGNIZE I study that explored the impact of long-term versus short-term outpatient SZC therapy on the rate of hospitalizations in patients with HK.19 In RECOGNIZE I, a lower proportion of patients with long-term SZC therapy (>3 months) than those with short-term SZC therapy (≤3 months) had HK-related hospitalizations (10.1% vs 15.1%; P < 0.05) and all-cause hospitalizations (22.5% vs 29.3%; P < 0.05).19 A follow-up study assessing HRU, the RECOGNIZE II study, also showed that patients receiving long-term outpatient SZC therapy (>3 months) incurred lower HK-related medical costs compared with patients not receiving SZC therapy.21 This study substantiates the results of the RECOGNIZE I study by conducting the analyses in a sample of patients matched on key characteristics using exact and propensity score matching. Although residual confounding was still possible, important confounders observed in the data were tightly controlled via the 2-step matching process.
The association of longer-term SZC use and lower acute HRU observed in this study is also consistent with prior studies evaluating the impact of other long-term potassium binder use on HRU and costs. In a retrospective claims-based study of US patiromer users (the VALUE study), long-term patiromer use (≥2 months) was associated with significantly lower all-cause medical costs than short-term use (<2 months), with the difference primarily driven by lower inpatient costs among long-term users.22 Another retrospective claims-based study of Japanese patients with HK found that chronic use of potassium binders (mean prescription duration: 1,151 days) was associated with fewer all-cause hospitalizations and adverse clinical outcomes, including kidney replacement therapy, compared with non-chronic potassium binder use (mean prescription duration: 115 days).23 These results suggest that HK management via continuous potassium-binding treatments may be associated with lower HRU and health care costs. In addition, the lower rates of HRU associated with long-term SZC use observed in this study may be due to patients experiencing fewer HK-related health events resulting from better HK management. However, as the reason for discontinuation of SZC and serum potassium values were not available in the claims data, the explicit reason why differences in HRU were observed was unable to be assessed in this study.
Longer-term SZC use was also associated with lower risk of HK-related hospitalizations and ED visits among patients using concomitant RAASi, possibly due to a lower risk of HK-related events or of RAASi discontinuation due to HK among long-term SZC users. In line with these speculations, the retrospective matched cohort OPTIMIZE II study showed that patients who experienced HK during RAASi therapy and received SZC were associated with lower HK-related medical costs than patients who discontinued RAASi after the HK diagnosis and did not receive SZC.24 Furthermore, in this sub-study among patients with RAASi use, long-term SZC users had a 34% lower rate of RAASi discontinuation over 6 months relative to short-term SZC users. Taken together, these results suggest that for patients receiving RAASi therapy, the long-term use of SZC to manage HK may improve RAASi persistence, which may in turn lower HK-related events and medical costs compared with short-term or no SZC use.
The potential benefits of long-term SZC use on HRU were similarly observed in patients with comorbid renal conditions such as CKD or kidney failure, which align with literature on improved clinical outcomes, including reduced risk of kidney failure or death, among long-term potassium binder users with HK and CKD.25 Prior studies have also shown that the majority of patients with CKD or kidney failure who used potassium binders to manage HK were able to continue their baseline RAASi therapy25,26 and maintain normal serum potassium levels.27 The current study similarly demonstrates that prolonged potassium binder use may also be associated with decreased HRU among patients with kidney disease or kidney failure, regardless of RAASi use.
Strengths and Limitations
Key strengths of this study included the tightly controlled double matching that minimized confounding between cohorts and the use of a large, closed claims database that allowed all encounters covered by the insurance providers to be visible. The study population thus constituted a broad group of real-world patients using SZC in an outpatient setting and not simply those who chose to participate in smaller research studies. Furthermore, the study population comprised comprehensive, stratified samples of clinically relevant patients, including RAASi users, patients with CKD, kidney failure, or HF, and patients with Medicare Advantage or Medicaid coverage, providing insights on the potential impact of outpatient SZC use duration on HRU in each of these patient subgroups.
One limitation of the study was that the duration of SZC therapy used for cohort assignment occurred during the follow-up period. Acute health events leading to health care encounters during follow-up could impact clinical management and affect the duration of SZC use. To reduce the risk of reverse causation, patients were censored on reinitiation of SZC after discontinuation or on initiation of a different potassium binder so that binder (re)initiation due to an acute health event resulting in a health care encounter would not contribute to cohort assignment. The study was also subject to limitations inherent to claims-based analyses. For instance, certain variables of interest (eg, mortality, treatment adherence, reasons for changes in dosing, and reasons for treatment discontinuation) were unavailable in claims data. Given the lack of information on reasons for discontinuation (eg, toxicity, adverse events, treatment preferences, and cost), it is possible that the short-term SZC cohort could include patients who discontinued SZC for reasons that may have impacted their HRU. Claims data also do not capture laboratory test results and electrocardiograph findings, which limited our ability to account for the confounding effect of HK severity on study outcomes. Furthermore, the lack of serum potassium values limits our ability to understand if changes in HRU are related to better potassium management or other reasons. Finally, patients in this study were required to have ≥7 months of continuous enrollment (6 months before index and ≥30 days of SZC supply after index). Thus, patients enrolled in their health plan for <7 months due to reasons such as a change in employment status were not included. If those patients differed from the overall population of patients initiating SZC, the results may not be generalizable to patients with <7 months of continuous enrollment.
In conclusion, in this exact-score and propensity score-matched RWE study of patients receiving outpatient SZC therapy, patients with long-term SZC use had significantly lower rates of HK-related hospitalizations or ED visits, HK-related hospitalizations, HK-related ED visits, and all-cause hospitalizations or ED visits than matched patients with short-term SZC use. The findings were consistent across patients using RAASi, patients with renal complications, patients with Medicare Advantage coverage, and patients with Medicaid coverage.
Article Information
Authors’ Full Names and Academic Degrees
Abiy Agiro, PhD, Erin E. Cook, ScD, Ali Greatsinger, MPH, Fan Mu, ScD, Jess Smith, MPH, Emily Reichert, MSc, Ellen Colman, PharmD, and Arun Malhotra, MD
Authors’ Contributions
Conceptualization, funding acquisition, investigation, methodology, project administration, resources, supervision, validation: AA, EEC, AG, FM, EC, and AM; conceptualization, data curation, formal analysis, investigation, methodology, resources, software, validation, visualization: JS. Each author contributed important intellectual content during article drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved.
Support
This study was funded by AstraZeneca. The study sponsor was involved in several aspects of the research, including the study design, interpretation of data, writing of the article, and the decision to submit the article for publication.
Financial Disclosure
Drs Agiro, Colman, and Malhotra are employees and stockholders of AstraZeneca. Drs Cook, Greatsinger, Mu, Smith, and Reichert are current employees of Analysis Group, Inc, a consulting firm that received payment from AstraZeneca for the development and conduct of this study and article.
Acknowledgements
The authors would like to thank Jingyi Chen, Manasvi Sundar, and Angela Zhao of Analysis Group, Inc for assistance with data analysis. Medical writing assistance was provided by professional medical writer, Flora Chik, PhD, MWC, an employee of Analysis Group, Inc, a consulting company that has provided paid consulting services to AstraZeneca, which funded the development and conduct of this study and article.
Prior Presentation
Part of the material in this article was presented at the National Kidney Foundation Spring Clinical Meetings held May 14-18, 2024, in Long Beach, CA, USA; the Heart Failure Society of America (HFSA) Annual Scientific Meeting held September 27-30, 2024, in Atlanta, GA, USA; the American Society of Nephrology Kidney Week held October 23-27, 2024, in San Diego, CA, USA; and the American Heart Association Annual Scientific Meeting held November 16-18, 2024, in Chicago, IL, USA, as poster presentations.
Data Sharing
The data that support the findings of this study are not available from the authors but are available with permission from HealthVerity. Restrictions apply to the availability of these data, which were used under license for this study.
Peer Review
Received July 28, 2025. Evaluated by 2 external peer reviewers, with direct editorial input from an Associate Editor and the Editor-in-Chief. Accepted in revised form December 22, 2025.
Footnotes
Complete author and article information provided before references.
Table S1: Baseline Characteristics in the CKD or Kidney Failure Sub-study.
Table S2: Baseline Characteristics in the Kidney Failure Sub-study.
Table S3: Baseline Characteristics in the HK Diagnosis Sub-study.
Table S4: Baseline Characteristics in the Medicare Sub-study.
Table S5: Baseline Characteristics in the Medicaid Sub-study.
Table S6: HRU Outcomes in the Kidney Failure or CKD Sub-study.
Table S7: HRU Outcomes in the Kidney Failure Sub-study.
Table S8: HRU Outcomes in the HK Diagnosis Sub-study.
Table S9: HRU Outcomes in the Medicare Advantage Sub-study.
Table S10: HRU Outcomes in the Medicaid Sub-study.
Supplementary Materials
Tables S1-S10.
References
- 1.Clase C.M., Carrero J.-J., Ellison D.H., et al. Potassium homeostasis and management of dyskalemia in kidney diseases: conclusions from a Kidney Disease: improving Global Outcomes (KDIGO) Controversies Conference. Kidney Int. January 2020;97(1):42–61. doi: 10.1016/j.kint.2019.09.018. [DOI] [PubMed] [Google Scholar]
- 2.Valdivielso J.M., Balafa O., Ekart R., et al. Hyperkalemia in chronic kidney disease in the new era of kidney protection therapies. Drugs. September 2021;81(13):1467–1489. doi: 10.1007/s40265-021-01555-5. [DOI] [PubMed] [Google Scholar]
- 3.Collins A.J., Pitt B., Reaven N., et al. Association of serum potassium with all-cause mortality in patients with and without heart failure, chronic kidney disease, and/or diabetes. Am J Nephrol. 2017;46(3):213–221. doi: 10.1159/000479802. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hunter R.W., Bailey M.A. Hyperkalemia: pathophysiology, risk factors and consequences. Nephrol Dial Transplant. 2019;34(suppl 3):iii2–iii11. doi: 10.1093/ndt/gfz206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kovesdy C.P., Matsushita K., Sang Y., et al. Serum potassium and adverse outcomes across the range of kidney function: a CKD Prognosis Consortium meta-analysis. Eur Heart J. May 1, 2018;39(17):1535–1542. doi: 10.1093/eurheartj/ehy100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Tanaka K., Saito H., Iwasaki T., et al. Association between serum potassium levels and adverse outcomes in chronic kidney disease: the Fukushima CKD cohort study. Clin Exp Nephrol. April 2021;25(4):410–417. doi: 10.1007/s10157-020-02010-7. [DOI] [PubMed] [Google Scholar]
- 7.Betts K.A., Woolley J.M., Mu F., McDonald E., Tang W., Wu E.Q. The prevalence of hyperkalemia in the United States. Curr Med Res Opin. June 2018;34(6):971–978. doi: 10.1080/03007995.2018.1433141. [DOI] [PubMed] [Google Scholar]
- 8.Fravel M.A., Meaney C.J., Noureddine L. Management of hyperkalemia in patients with chronic kidney disease using renin angiotensin aldosterone system inhibitors. Curr Hypertens Rep. November 2023;25(11):395–404. doi: 10.1007/s11906-023-01265-1. [DOI] [PubMed] [Google Scholar]
- 9.Palmer B.F., Carrero J.J., Clegg D.J., et al. Clinical management of hyperkalemia. Mayo Clin Proc. March 2021;96(3):744–762. doi: 10.1016/j.mayocp.2020.06.014. [DOI] [PubMed] [Google Scholar]
- 10.Fried L.F., Emanuele N., Zhang J.H., et al. Combined angiotensin inhibition for the treatment of diabetic nephropathy. N Engl J Med. November 14, 2013;369(20):1892–1903. doi: 10.1056/NEJMoa1303154. [DOI] [PubMed] [Google Scholar]
- 11.Weinberg J.M., Appel L.J., Bakris G., et al. Risk of hyperkalemia in nondiabetic patients with chronic kidney disease receiving antihypertensive therapy. Arch Intern Med. September 28, 2009;169(17):1587–1594. doi: 10.1001/archinternmed.2009.284. [DOI] [PubMed] [Google Scholar]
- 12.Davis J., Israni R., Betts K.A., et al. Real-World management of hyperkalemia in the emergency department: an electronic medical record analysis. Adv Ther. February 2022;39(2):1033–1044. doi: 10.1007/s12325-021-02017-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Press release: Lokelma approved in the US for the treatment of adults with hyperkalaemia; May 18, 2018. Astra Zeneca. https://www.astrazeneca.com/media-centre/press-releases/2018/lokelma-approved-in-the-us-for-the-treatment-of-adults-with-hyperkalaemia-21052018.html#
- 14.Betts K.A., Woolley J.M., Mu F., Xiang C., Tang W., Wu E.Q. The cost of hyperkalemia in the United States. Kidney Int Rep. March 2018;3(2):385–393. doi: 10.1016/j.ekir.2017.11.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Desai N.R., Reed P., Alvarez P.J., Fogli J., Woods S.D., Owens M.K. The economic implications of hyperkalemia in a Medicaid managed care population. Am Health Drug Benefits. November 2019;12(7):352–361. [PMC free article] [PubMed] [Google Scholar]
- 16.Fitch K., Woolley J.M., Engel T., Blumen H. The clinical and economic burden of hyperkalemia on Medicare and commercial payers. Am Health Drug Benefits. June 2017;10(4):202–210. [PMC free article] [PubMed] [Google Scholar]
- 17.Mu F., Betts K.A., Woolley J.M., et al. Prevalence and economic burden of hyperkalemia in the United States Medicare population. Curr Med Res Opin. August 2020;36(8):1333–1341. doi: 10.1080/03007995.2020.1775072. [DOI] [PubMed] [Google Scholar]
- 18.Dashputre A.A., Gatwood J., Sumida K., et al. Association of dyskalemias with short-term health care utilization in patients with advanced CKD. J Manag Care Spec Pharm. October 2021;27(10):1403–1415. doi: 10.18553/jmcp.2021.27.10.1403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Pollack C.V., Jr., Agiro A., Mu F., et al. Impact on hospitalizations of long-term versus short-term therapy with sodium zirconium cyclosilicate during routine outpatient care of patients with hyperkalemia: the recognize I study. Expert Rev Pharmacoecon Outcomes Res. February 2023;23(2):241–250. doi: 10.1080/14737167.2023.2161514. [DOI] [PubMed] [Google Scholar]
- 20.Cohen J. 2nd ed. Lawrence Erlbaum Associates; 1988. Statistical Power Analysis for the Behavioral Sciences. [Google Scholar]
- 21.Agiro A., Dwyer J.P., Oluwatosin Y., Desai P. Medical costs in patients with hyperkalemia on long-term sodium zirconium cyclosilicate therapy: the RECOGNIZE II study. Clinicoecon Outcomes Res. 2023;15:691–702. doi: 10.2147/CEOR.S420217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Gupta A., Kammerer J., Shaik I., Mukherjee K.G., Oliveira J., Thakar C. Evaluation of longer- vs short-term use of patiromer on health care resource utilization in the patiromer longer-term use evaluation (VALUE) study. J Manag Care Spec Pharm. January 2024;30(1):52–60. doi: 10.18553/jmcp.2023.23100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kanda E., Morita N., Yajima T. Impact of chronic potassium binder treatment on the clinical outcomes in patients with hyperkalemia: results of a nationwide hospital-based cohort study. Front Physiol. 2023;14 doi: 10.3389/fphys.2023.1156289. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Dwyer J.P., Agiro A., Desai P., Oluwatosin Y. Impact of sodium zirconium cyclosilicate plus renin-angiotensin-aldosterone system inhibitor therapy on short-term medical costs in hyperkalemia: OPTIMIZE II real-world study. Adv Ther. November 2023;40(11):4777–4791. doi: 10.1007/s12325-023-02631-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Obi Y., Thomas F., Dashputre A.A., Goedecke P., Kovesdy C.P. Long-term patiromer use and outcomes among US Veterans with hyperkalemia and CKD: A propensity-matched cohort study. Kidney Med. January 2024;6(1) doi: 10.1016/j.xkme.2023.100757. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Riccio E., D’Ercole A., Sannino A., et al. Real-world management of chronic and postprandial hyperkalemia in CKD patients treated with patiromer: A single-center retrospective study. J Nephrol. May 2024;37(4):1077–1084. doi: 10.1007/s40620-024-01897-9. [DOI] [PubMed] [Google Scholar]
- 27.Costa D., Patella G., Provenzano M., et al. Hyperkalemia in CKD: an overview of available therapeutic strategies. Front Med (Lausanne) 2023;10 doi: 10.3389/fmed.2023.1178140. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Tables S1-S10.