Hyperkalemia burden and treatment patterns in Chinese patients on hemodialysis: final analysis of a prospective multicenter cohort study (PRECEDE-K)

Abstract

Objectives

Patients with end-stage renal disease (ESRD) on hemodialysis (HD) are at risk for hyperkalemia (HK), associated with cardiac arrhythmia and sudden death. Data on the burden of HK and management techniques among HD patients in China are still scarce. This study assessed the treatment modalities, recurrence, and prevalence of HK in Chinese HD patients.

Methods

In this prospective cohort study conducted from May 2021 to July 2022, patients aged ≥18 years who had ESRD and were on HD were enrolled from 15 centers in China (up to 6 months).

Results

Overall, 600 patients were enrolled. At the baseline visit, mean (± standard deviation) urea reduction ratio was 68.0% ± 9.70 and Kt/V was 1.45 ± 0.496. Over 6 months, 453 (75.5%) patients experienced HK, of whom 356 (78.6%) recurred. Within 1, 2, 3, 4, 5, and 6 months, 203 (44.8%), 262 (57.8%), 300 (66.2%), 326 (72.0%), 347 (76.6%), and 356 (78.6%) patients had at least one HK recurrence event, respectively. The proportions of patients with ≥1, 2, 3, 4, 5, or 6 HK recurrence events were 356 (78.6%), 306 (67.5%), 250 (55.2%), 208 (45.9%), 161 (35.5%), and 110 (24.3%), respectively. Among the 453 patients who experienced HK, only 24 (5.3%) were treated with potassium binders: seven (1.5%) with sodium polystyrene sulfonate, 13 (2.9%) with calcium polystyrene sulfonate, and six (1.3%) with sodium zirconium cyclosilicate.

Conclusion

Since HK is a chronic illness, long-term care is necessary. Patients on HD should have effective potassium management on non-dialysis days, yet our real-world population rarely used potassium binders.

Trial registration

ClinicalTrials.gov Identifier NCT04799067

Introduction

Hyperkalemia (HK) is a common complication of ESRD, arising due to the kidneys’ decreased ability to excrete potassium [1,2]. HK can cause cardiac dysrhythmia and sudden cardiac death, and is associated with worse clinical outcomes, including increased mortality and hospitalization rates [1–4]. Hemodialysis (HD) is the most common mode of renal replacement therapy (RRT) for patients with ESRD worldwide [5,6]. One of its goals entail the removal of excess serum potassium (sK) to control for HK [7].

Despite HD, HK prevalence remains high among patients with ESRD as many of them have suboptimal control of their sK levels [2,8–10]. This arises due to the intermittent nature of HD where patients have sK rebound during the interdialytic periods [11,12]. HK among patients on HD is observed to be persistent and recurrent, with up to 80% of patients experiencing recurrent events within 3 months of an initial episode [8,10]. Moreover, patients with HK commonly experience large fluctuations in sK levels during HD [13], which has been shown to increase the risk of cardiac dysrhythmia and mortality [14–16].

To date, there is a paucity of data on HK burden and treatment patterns among patients receiving HD in China. While maintenance of sK at normal levels over the long term is recognized as an important strategy to improve the prognosis in this group of patients, there is a lack of updated local clinical guidelines based on recent clinical trial results. The PRECEDE-K study (ClinicalTrials.gov Identifier: NCT04799067) aimed to characterize the prevalence, recurrence, and treatment patterns of HK among Chinese patients receiving HD. An interim analysis has previously reported baseline characteristics of the PRECEDE-K cohort, as well as HK prevalence, serum–dialysate potassium gradient, and sK fluctuation patterns at the baseline HD visit [17]. At that visit, 39.6% of patients had pre-dialysis HK, and most patients (97.7%) were prescribed a dialysate potassium concentration of 2.0 mmol/L. As a result, over 40% of patients had a serum-dialysate potassium gradient exceeding 3 mmol/L, and an intradialytic sK reduction of ≥1 mmol/L occurred in over 75% of the patients. These findings indicate that pre-dialysis HK and sK fluctuations over the HD cycles occur frequently among Chinese patients on HD. Here, we report the final analysis of PRECEDE-K, including HK prevalence, recurrence, associated risk factors, and treatment patterns, after 6 months of follow-up.

Materials and methods

Study design and participants

This prospective, observational cohort study was conducted from 17 May 2021 to 18 July 2022. Patients aged ≥18 years with ESRD who were on HD treatment across 15 HD centers from tertiary and secondary hospitals in China were enrolled. To minimize selection bias, eligible patients were enrolled consecutively without any preferences from the investigators. Key exclusion criteria included acute kidney injury; expected renal transplantation in the next 6 months; any of intracranial hemorrhage, elevated intracranial pressure, or shock that cannot be corrected by drugs within one month before enrollment; failure to establish vascular access; peritoneal dialysis; or otherwise, unsuitable for the study as deemed by the investigators.

The study protocol and interim analysis results have been previously published [17,18]. Patients received HD at frequencies of two-times weekly (minimum) to three-times weekly (maximum). In a week, two-times weekly HD gives rise to two long interdialytic intervals (LIDIs), defined as ≥2 days interval between HD sessions, whereas three-times weekly HD results in one LIDI and two short interdialytic intervals (SIDIs), defined as a 1-day interval between HD sessions. At the baseline HD visit (Visit 1), patients were in a LIDI. Subsequently, patients were followed up during a LIDI every month for up to 6 months, or until death, loss to follow-up, change of RRT modality, informed consent form (ICF) withdrawal, or termination as deemed necessary by investigators, whichever occurred earlier. Patients who received three-times weekly HD had an additional follow-up visit at Day 3 or 5 during a SIDI.

At Visit 1, demographic characteristics, medical history, ESRD etiology, concomitant medications, electrocardiogram, and pre- and post-dialysis sK and other laboratory measurements were collected. At each follow-up LIDI visit, pre-dialysis sK measurements (post-dialysis sK measurements were not mandatory beyond Visit 1), dialysis parameters (dialysis adequacy, the prescribed duration of dialysis, and dialysate potassium concentration), concomitant medications, and other clinical and laboratory findings were obtained. For patients with a SIDI visit, pre- and post-dialysis sK measurements were obtained.

This study was conducted in accordance with the Declaration of Helsinki, the International Conference on Harmonization’s Guidelines for Good Clinical Practice, and applicable local legislation on non-interventional studies and/or observational studies. All patients provided written and oral ICFs pre-enrollment. The protocol and all its amendments were approved by the Shanghai Jiao Tong University School of Medicine Renji Hospital Ethics Committee (2020-040) and the ethics committee of each participating center. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.

Outcomes

The primary endpoint was the proportion of patients who experienced any pre-dialysis HK event, defined as sK >5.0 mmol/L [19], at study enrollment or during the 6-month follow-up. Secondary endpoints were the proportions of patients who experienced HK recurrence within 1, 2, 3, 4, 5, or 6 months after a first event on study; the proportions of patients with ≥1, 2, 3, 4, 5, or 6 HK recurrence events; intradialytic sK shift, defined as the difference between pre- and post-dialysis sK levels (last measurement before HD for pre-dialysis sK and the first measurement after HD for post-dialysis sK if multiple were taken) during the first week (results previously published [17]); sK levels at LIDI and SIDI during the first week in patients who received HD three times a week; the proportion of patients with HK, as well as the proportion of HK events, treated with any potassium binders, including sodium polystyrene sulfonate (SPS), calcium polystyrene sulfonate (CPS) or sodium zirconium cyclosilicate (SZC); mean daily doses of SPS, CPS or SZC and the durations of treatment in patients treated with any potassium binder. Exploratory endpoints were risk factors associated with HK occurrence and recurrence.

Statistical analysis

As previously described, a sample size of 600 patients was planned to provide a precision (half width of 95% confidence interval [CI]) estimate of 3.5%–3.9% for the primary endpoint [18]. The full analysis set (FAS), which was used for all analyses, included all enrolled patients. Missing data were not imputed unless otherwise specified.

The analysis was primarily descriptive. Categorical data were shown as numbers and percentages (based on non-missing data unless otherwise specified). Continuous data were presented as means with standard deviation (SD), and medians with interquartile ranges and ranges. Two-sided 95% CIs for patient proportions were calculated using the Wald simple asymptotic method in a post hoc, exploratory manner. Sensitivity analyses for HK prevalence and recurrence were also performed using higher thresholds of sK >5.3 and 5.5 mmol/L for HK.

Potential risk factors of HK occurrence or recurrence, such as patient demographics, dialysis parameters, and medical history of special interest, were included in the univariate logistic regression model as independent variables; and those presenting with p ≤ 0.1 were entered in the multivariate logistic regression model in a stepwise manner. Unadjusted and adjusted odds ratio indicating the association of each risk factor with HK were determined, along with 95% CI estimation and two-sided p-values.

Results

Patient disposition and baseline characteristics

A total of 604 patients were screened, among whom 600 were eligible and enrolled into the FAS; 545 patients (90.8%) from the FAS completed the study. The patient disposition is shown in Supplementary Figure S1. No patients were excluded from the FAS due to protocol deviations.

Patients’ characteristics at baseline are summarized in Table 1 and were previously described [17]. Median age was 55.0 years, and 403 (67.2%) patients were male. Prior HK (in the past 6 months) was present in 60.0% of patients; other common medical histories included hypertension (81.2%), renal anemia (81.2%), and hyperphosphatemia (46.2%). Only 2.5% of patients had used a potassium binder at baseline. Most patients presented with abnormalities in clinical chemistry measurements at baseline; however, no clinically important trends were observed in the mean changes from baseline for clinical chemistry and blood gas analyses throughout the study. Results from these laboratory evaluations by visit are presented in Supplementary Table S1. Concomitant medication included renin-angiotensin system inhibitors (RASi) (37% of patients) and diuretics (6.8%). The commonly used RASi were angiotensin receptor blockers like irbesartan (5.0%), valsartan (2.7%), and telmisartan (1.5%); angiotensin-converting enzyme such as enalapril maleate (0.5%), fosinopril sodium (0.2%), perindopril arginine (0.2%). Furthermore, widely used combination of RASi included sacubitril:valsartan (10.7%), and amlodipine besilate:valsartan (8.7%).

Table 1.

Baseline characteristics in the FAS.

Parameter	FASa (N = 600)
Demographics
Age, years	n = 600
Mean (SD)	54.3 (12.9)
Median (IQR)	55.0 (45.0–63.5)
Sex, n (%)	n = 600
Male	403 (67.2)
Disease characteristics and concomitant conditions
Etiology of ESRD, n (%)	n = 485
Diabetic kidney disease	135 (27.8)
Primary glomerulonephritis	188 (38.8)
Hypertensive renal disease	77 (15.9)
Others	91 (18.8)
Medical history, n (%)	N = 600
Presenta	599 (99.8)
Hyperkalemia	360 (60.0)
Renal anemia	490 (81.2)
Hypertension	487 (81.2)
Hyperphosphatemia	283 (47.2)
Diabetes mellitus	119 (19.8)
Metabolic acidosis	79 (13.2)
Hyperlipidemia	75 (12.5)
Coronary artery disease	78 (13.0)
Hyperuricemia	61 (10.2)
Dialysis parameters
Type of hemodialysis, n (%)	n = 599
Hemodialysis	529 (88.3)
Hemodiafiltration	60 (10.0)
Hemoperfusion combined with hemodialysis or hemodiafiltration	10 (1.7)
Vascular access, n (%)	n = 599
Arteriovenous graft	11 (1.8)
Central tunneled dialysis catheter	17 (2.8)
Arteriovenous fistula	567 (94.7)
Other (temporary catheter)	4 (0.7)
Dialysis frequency, n (%)	n = 599
Three-times weekly	509 (85.0)
Two-times weekly	35 (5.8)
Five times every 2 weeks	55 (9.2)
Dialysate potassium concentration, mmol/L	n = 599
Mean (SD)	2.0 (0.1)
Median (IQR)b	2.0 (2.0–2.0)
Hemodialysis duration, hours	n = 599
Mean (SD)	4.0 (0.2)
Median (IQR)	4.0 (4.0–4.0)
Urea reduction ratio, %	n = 589
Mean (SD)	68.0 (9.70)
Median (IQR)	70.0 (60.0–70.0)
Kt/V	n = 577
Mean (SD)	1.45 (0.496)
Median (IQR)	1.38 (1.20–1.61)
Gradient between dialysate and serum potassium (mmol/L), n (%)	n = 596
0–<1	2 (0.3)
1–<2	81 (13.6)
2–<3	269 (45.1)
3–<4	213 (35.7)
4–<5	25 (4.2)
5–<6	6 (1.0)
Hyperkalemia treatment
Use of potassium binders, n (%)	n = 600
Yes	15 (2.5)
No	585 (97.5)

^a Medical history that occurred in ≥10% of patients are presented.

^bRange: 2.0–3.0 mmol/L.

ESRD, end-stage renal disease; FAS, full analysis set; IQR, interquartile range; SD, standard deviation.

Dialysis parameters by visit

At Visit 1, the most common type of dialysis in the cohort was conventional HD alone (88.3%), and the HD duration in majority of the patients was 4 h. Most patients received dialysis three-times weekly (85.0%). The majority of patients (97.7%) were prescribed dialysate potassium concentration of 2.0 mmol/L, and the serum–dialysate potassium gradient was greater than 3 mmol/L in 40.9% of patients. Most patients had adequate dialysis: the mean (± SD) urea reduction ratio was 68.0% ± 9.70, and the mean (± SD) Kt/V was 1.45 ± 0.496. All dialysis parameters remained relatively consistent until the end of study. Full information of the dialysis parameters by visit is presented in Supplementary Table S2.

HK prevalence

A total of 453 (75.5%; 95% CI, 72.1–78.9) patients in the FAS experienced any HK event over 6 months (Figure 1a). sK was >5.3, 5.5, 6.0, 6.5, and 7.0 mmol/L in 357 (59.5%), 297 (49.5%), 171 (28.5%), 83 (13.8%), 28 (4.7%), and 9 (1.5%) patients, respectively.

Figure 1.

(a) Prevalence of HK occurrence; (b) Frequency of HK recurrence; (c) Time to HK recurrence. ^aIn patients who had a first HK event. HK, hyperkalemia; sK, serum potassium.

HK recurrence

Among 453 patients who had any HK event, 356 (78.6%; 95% CI, 74.8–82.4) patients had at least one HK recurrence event over the 6-month follow-up period. Within 1, 2, 3, 4, 5, and 6 months, 203 (44.8%), 262 (57.8%), 300 (66.2%), 326 (72.0%), 347 (76.6%), and 356 (78.6%) patients had at least one HK recurrence event, respectively (Figure 1b). In the sensitivity analysis, the proportions were 110 (24.3%), 152 (33.6%), 186 (41.1%), 218 (48.1%), 249 (55.0%), and 273 (60.3%), respectively, when HK was defined as sK >5.3 mmol/L; and 75 (16.6%), 111 (24.5%), 137 (30.2%), 160 (35.3%), 185 (39.7%), and 202 (44.6%) when HK was defined as sK >5.5 mmol/L.

The proportions of patients with ≥1, 2, 3, 4, 5, or 6 HK recurrence events were 356 (78.6%), 306 (67.5%), 250 (55.2%), 208 (45.9%), 161 (35.5%), and 110 (24.3%), respectively (Figure 1c). In the sensitivity analysis, the proportions were 273 (60.3%), 200 (44.2%), 152 (33.6%), 108 (23.8%), 77 (17.0%), and 45 (9.9%), respectively, when HK was defined as sK >5.3 mmol/L; and 202 (44.6%), 140 (30.9%), 102 (22.5%), 74 (16.3%), 42 (9.3%), and 25 (5.5%) when HK was defined as sK >5.5 mmol/L.

Difference in pre-dialysis sK following a LIDI versus a SIDI

In patients who received HD three-times weekly, the mean (SD) pre-dialysis sK following a LIDI and a SIDI during the first week was 4.8 (0.7) and 4.7 (0.7) mmol/L, respectively, yielding a mean (SD) sK shift of −0.1 (0.4) mmol/L from LIDI to SIDI. Pre-dialysis HK was observed in 187 (36.9%) patients after a LIDI and 146 (31.0%) patients after a SIDI (Table 2).

Table 2.

Pre-dialysis sK levels following a LIDI vs a SIDI during the first week in patients who received HD three-times weekly.

Pre-dialysis serum potassium (mmol/L)	n = 509a
Mean (SD)
LIDI	n = 507
Mean (SD)	4.79 (0.73)
Median (range)	4.75 (2.9–7.9)
sK (mmol/L), n (%)
>5.0	187 (36.9)
>5.5	77 (15.2)
>6.0	24 (4.7)
>6.5	11 (2.2)
SIDI	n = 471
Mean (SD)	4.70 (0.69)
Median (range)	4.65 (3.2–7.4)
sK (mmol/L), n (%)
>5.0	146 (31.0)
>5.5	60 (12.7)
>6.0	17 (3.6)
>6.5	6 (1.3)
Difference (LIDI vs SIDI)	n = 469
Mean (SD)	–0.10 (0.43)

HD, hemodialysis; LIDI, long interdialytic interval; SD, standard deviation; SIDI, short interdialytic interval.

Use of potassium binders for HK treatment

Among the 453 patients who experienced HK over the 6-month follow-up period, 24 (5.3%) were treated with a potassium binder: 7 (1.6%) with SPS, 13 (2.9%) with CPS, and 6 (1.3%) with SZC. Out of the 2065 HK events, 54 (2.6%) were treated with potassium binders: 17 (0.8%) with SPS, 28 (1.4%) with CPS, and 13 (0.6%) with SZC. The mean (SD) daily doses of SPS, CPS, and SZC in grams were 16.9 (7.3), 11.8 (9.8), and 10.7 (8.9), respectively; the mean (SD) treatment durations in days were 19.4 (25.4), 26.0 (43.3), and 4.7 (1.5), respectively. Full data on the use of potassium binders for HK treatment in the cohort are shown in Table 3 and Supplementary Table S3.

Table 3.

Summary of potassium binder usage in patients with any HK event.

Potassium binder usage	n = 453
Total number of HK events	2065
Treated with potassium binders, n (% [95%CI])	54 (2.62 [1.93–3.3])
SPS	17 (0.82 [0.43–1.21])
CPS	28 (1.36 [0.86–1.85])
SZC	13 (0.63 [0.29–0.97])
Total number of patients with HK	453
Treated with potassium binders, n (% [95% CI])	24 (5.30 [3.24–7.36])
SPS	7 (1.55 [0.41–2.68])
CPS	13 (2.87 [1.33–4.41])
SZC	6 (1.32 [0.27–2.38])

CI, confidence interval; CPS, calcium polystyrene sulfonate; HK, hyperkalemia; SD, standard deviation; SPS, sodium polystyrene sulfonate; SZC, sodium zirconium cyclosilicate.

Risk factors for HK occurrence and recurrence

Based on the univariate logistic regression, risk factors identified as independent risk factors for HK occurrence were gender, dialysis vintage, baseline dialysis frequency, baseline ultrafiltration rate, history of HK, and history of diabetes. Multivariable analyses found that patients who were of female gender, or those with a longer dialysis vintage, a lower dialysis frequency, or a history of HK were significantly more likely to have HK (Table 4).

Table 4.

Risk factors for hyperkalemia occurrence and recurrence.

HK occurrence
Risk factor	Univariate analysis			Multivariate analysis
	n	OR [95% CI]	p-value	n	OR [95% CI]	p-value
Sex (Female vs male)	600	1.70 [1.11–2.59]	0.0138	592	1.90 [1.18–3.05]	0.0079
Age (per increase of 1 year)	600	1.00 [0.99–1.02]	0.9346	–
Dialysis vintage (per increase of 1 year)	592	0.96 [0.93–1.00]	0.0324	592	0.95 [0.91–0.99]	0.0080
Baseline dialysis frequency
Two-times weekly vs three-times weekly	544	1.23 [0.55–2.77]	0.6170	592	2.61 [1.55–4.39]a	0.0003
Five times per 2 weeks vs three-times weekly	564	9.65 [2.32–40.12]	0.0018
Baseline dialysate potassium concentration (per increase of 1 mmol/L)	599	0.66 [0.19–2.29]	0.5148	–
Baseline ultrafiltration rate (per increase of 1 ml/h)	599	1.00 [1.00–1.00]	0.0013	592	1.00 [1.00–1.00]	0.0004
Baseline urea reduction ratio (per increase of 1 unit)	589	0.59 [0.09–4.05]	0.5926	–
Baseline Kt/V (per increase of 1 unit)	577	0.91 [0.63–1.32]	0.6283	–
24-hour urine volume (per increase of 1 mL)b	90	1.00 [1.00–1.00]	0.2015	–
Atherosclerotic heart disease (Yes vs No)	600	–		–
Congestive heart failure (Yes vs No)	600	–		–
Diabetes (Yes vs No)	600	1.46 [0.96–2.22]	0.0741	–
Hypertension (Yes vs No)	600	1.45 [0.69–3.03]	0.3282	–
Hyperkalemia (Yes vs No)	600	4.96 [3.32–7.41]	<0.0001	592	5.12 [3.36–7.79]	<0.0001
Baseline potassium binder treatment (Yes vs No)	600	4.66 [0.61–35.71]	0.1389	–
Sex (Female vs male)	600	1.42 [0.99–2.03]	0.0541	–
Age (per increase of 1 year)	600	0.99 [0.98–1.01]	0.2975	–
Dialysis vintage (per increase of 1 year)	592	0.97 [0.94–1.00]	0.0664	–
Baseline dialysis frequency
Two-times weekly vs three-times weekly	544	1.54 [0.74–3.22]	0.2462	87	3.01 (1.18–7.67)a	0.0210
Five times per 2 weeks vs three-times weekly	564	2.83 [1.43–5.61]	0.0028
Baseline dialysate potassium concentration (per increase of 1 mmol/L)	599	0.56 [0.18–1.79]	0.3277	–
Baseline ultrafiltration rate (per increase of 1 ml/h)	599	1.00 [1.00–1.00]	<0.0001	–
Baseline urea reduction ratio (per increase of 1 unit)	589	0.19 [0.03–1.06]	0.0576	–
Baseline Kt/V (per increase of 1 unit)	577	0.82 [0.58–1.14]	0.2344	–
24-hour urine volume (per increase of 1 mL)b	90	1.00 [1.00–1.00]	0.0415	87	1.00 [1.00–1.00]	0.0110
Atherosclerotic heart disease (Yes vs No)c	600	–		–
Congestive heart failure (Yes vs No)c	600	–		–
Diabetes (Yes vs No)	600	1.51 [1.05–2.17]	0.0252	87	2.71 [0.92–7.98]	0.0710
Hypertension (Yes vs No)	600	1.93 [0.97–3.84]	0.0598
Hyperkalemia (Yes vs No)	600	4.50 [3.17–6.40]	<0.0001	87	6.07 [2.01–18.29]	0.0014
Baseline potassium binder treatment (Yes vs No)	600	2.61 [0.73–9.35]	0.1406	–

^a Two-times weekly or five times per 2 weeks vs three-times weekly.

^b 24-h urine volume: the mean volume was used if a patient had urine volume measured on both visits; otherwise, the last available measurement was used.

^cOnly one patient was assessed with echocardiography at baseline.

CI, confidence interval; NA, not applicable; OR, odds ratio.

Based on the univariate logistic regression, risk factors identified as independent risk factors for HK recurrence were gender, dialysis vintage, baseline dialysis frequency, baseline ultrafiltration rate, baseline urea reduction ratio, baseline 24-h urine volume, history of diabetes, history of hypertension, and history of HK. Multivariable analyses found that patients with a lower dialysis frequency, history of diabetes, and history of HK were significantly more likely to have HK recurrence (Table 4).

Discussion

To our knowledge, PRECEDE-K is the first and largest prospective cohort study that investigated the prevalence, recurrence, and treatment patterns of HK in Chinese patients with ESRD on HD. A strength of this study is the prospective design that allows for regular sK evaluations to monitor HK recurrence in this real-world cohort of patients over time. Over a 6-month follow-up, we found a high overall HK prevalence of 75.5%. HK recurrence occurred in 78.6% of patients within this short follow-up timespan; 68% and 36% of patients had ≥2 and ≥5 recurrence events, respectively, and 45% and 66% of patients recurred within 1 and 3 months. These observations are consistent with the disease pattern of chronic HK, which requires ongoing management to correct the underlying disturbances in potassium balance [20]. Risk factors that were significantly associated with HK occurrence and/or recurrence included a lower dialysis frequency and a history of HK. In our study, dialysis frequency and the history of HK were identified as potential risk factors for the occurrence and recurrence of HK. This can be attributed to the discharge of larger amounts of potassium upon frequent dialysis [21–23]. Among patients who received three-times weekly HD, pre-dialysis HK occurred in 36.9% of patients at LIDI versus 31.0% at SIDI. Potassium binders were infrequently used in only 5.3% of the patients with HK; 4.4% received traditional potassium binders and 1.3% received SZC. The Standards document considers both Kt/V and urea reduction ratio as indicators of the adequacy of hemodialysis. Based on these data the standard value for urea reduction ratio (>65%) and Kt/V (>1.2%) [24,25].

The prospective monthly sK test and follow-up allowed for accurate prediction of pre-dialysis HK rate in the PRECEDE-K cohort, which was found to be higher compared with previous reports. While the 6-month HK occurrence rate was 75.5% in the PRECEDE-K cohort, previous studies reported 1-year rates of 73.8% (prospective study in France, sK > 5.1 mmol/L) and 73.9% (retrospective study in the US, sK > 5.0 mmol/L) [8,10,26]. In the prospective DOPPS study, HK (based on peak sK) was observed in 75% of the total 4-monthly periods evaluated for China [2]. Our observed HK recurrence rate of 78.6% within 6 months was also considerably high among other global observational studies. This high recurrence rate can be attributed to the frequent sK test conducted once every month in our study. A study found that a higher frequency of sK testing leads to a higher HK recurrence rate [22]. In a retrospective US cohort study, 59.5%, 47.7%, and 38.6% of the patients receiving HD three-times weekly had ≥2, 3, or 4 HK events (defined as sK >5.0 mmol/L) within a 1-year period [8]. Although a high HK recurrence rate of 73.2% within 3 months (defined as sK > 5.1 mmol/L) was observed in a French prospective cohort, this might be due to a larger proportion of patients (47.7%) in the cohort who received a high dialysate potassium concentration of 3.0 mmol/L compared with only 1.8% in the present study [10].

Consistent with the high rates of HK recurrence observed in this study and previous reports, we identified prior HK as a risk factor for HK events [10]. Collectively, this evidence suggests that patients with prior HK are highly susceptible to subsequent events. Another risk factor identified for HK events was a lower dialysis frequency. This may be explained by the intermittent nature of potassium removal in HD contributing to sK accumulation and rebound at interdialytic intervals, especially at LIDI compared with SIDI [3,27]. In a retrospective US study, HK prevalence (defined as sK ≥5.5 mmol/l) was 2.0–2.4 times higher on the day after a LIDI than on the day after a SIDI [3]. Another retrospective US study showed that the proportion of patients with pre-dialysis sK ≥5.0 mmol/L was slightly higher after a LIDI (49.0% on Monday) versus after a SIDI (42.7% on Wednesday and 43.6% on Friday) [3]. In the present study, we observed a modest 0.1 mmol/L increase in mean pre-dialysis sK following a LIDI versus SIDI and a slightly higher proportion of patients with HK after a LIDI (36.9%) versus SIDI (31.0%). In contrast to the results from previous retrospective studies, the subtle difference in the HK prevalence following a LIDI versus a SIDI in this study may suggest that HK remains a potential problem during SIDI, a finding that was not captured in retrospective studies and warrants further confirmation.

The risk factors for HK occurrence among patients receiving HD differed between studies. Other risk factors previously reported to be associated with HK included age <40 years (versus ≥80 years), a history of diabetes, female sex, concomitant renin-angiotensin-aldosterone system inhibitor medication, HD sessions <4 h, and the use of a high dialysate potassium concentration [8,23,26,28]. Further research should be conducted in the Asian population to assess the validity of the risk factors for HK occurrence and recurrence identified in this study. These risk factors could potentially be useful in identifying patient populations at high risk of HK in whom close monitoring is required for timely management.

The association between HK and poor outcomes has been widely reported among patients with ESRD receiving HD [3,27,29–32] including a 2.7-fold increased risk of sudden cardiac death [4]. HK is also a risk factor for sK fluctuations in patients on HD [15], which is associated with 42% higher risk of all-cause death and 83% higher risk of cardiovascular death [14–16]. With the high rates of HK occurrence and recurrence, as well as considerably large intradialytic shifts observed in the PRECEDE-K cohort, effective management of HK to maintain sK levels within a normal range during the interdialytic period is required [3,27,29–32]. The high hyperkalemia recurrence rate and the huge intradialytic sK shift will deeply prove the importance of HK management on the non-dialysis day, not only relying on the hemodialysis on dialysis day [17].

Using non-pharmacological methods alone, such as dietary restrictions, may be insufficient for management of HK; in a cohort of over 8000 adults on HD, dietary potassium intake was not associated with sK levels, as well as all-cause and cardiovascular mortalities [33]. The use of novel potassium binders showed efficacy in managing HK among patients on HD [34]. SZC is the only drug in this class that has been indicated in patients on chronic HD, and the only novel potassium binder approved for use in China to date. In the Phase IIIb DIALIZE trial (NCT03303521), which evaluated patients with ESRD who received three-times weekly HD with pre-dialysis HK (n = 196), 41.6% of the patients in the SZC arm were responders (defined as patients who maintained sK levels of 4.0–5.0 mmol/L for at least three of four LIDIs without using rescue therapy), compared with 1.0% in the placebo arm (odds ratio 68.8; 95% CI, 10.9–2810; p < 0.001) [34]. The patients who received placebo experienced large shifts in intradialytic sK, and their mean pre-dialysis sK levels consistently remained at ≥5.0 mmol/L, with a mean (SD) pre-dialysis sK of 5.8 (0.7) mmol/L at the end of the study [13]. A similar trend was observed in patients with severe HK (sK ≥6.0 mmol/L) at baseline, where 34.8% of patients in the SZC group responded versus 0% in the placebo group (p < 0.0001) [35]. Furthermore, the improved control of HK with SZC treatment lowered the serum-dialysate potassium gradient compared with placebo (mean difference: −0.74 mmol/L after 57 days) and the potassium gradient was maintained or reduced in more than 80% of patients, without the need for change in dialysate potassium concentration [36]. This could potentially reduce the risk of hospitalization and emergency department visits associated with steep serum-dialysate potassium concentration [27]. The efficacy and safety of SZC in managing HK in the Chinese population was similarly reported in the DIALIZE China trial (NCT04217590) [37].

While the efficacy of the novel potassium binder SZC in controlling for HK and limiting sK fluctuations on non-dialysis days has been demonstrated in the DIALIZE trial, its use in our real-world cohort of Chinese patients on HD remains limited at 0.9%. To overcome chronic HK and potassium fluctuations in patients on HD, sK management on non-dialysis days with oral potassium binders should be encouraged, which has the potential to improve clinical outcomes. The more frequent dialysis frequency will discharge more potassium and

Limitations

Our study has a few limitations. Firstly, HK prevalence and recurrence were evaluated over a relatively short period of time of 6 months. Secondly, our study did not assess the association between HK and clinical outcomes, such as cardiovascular mortality, which need long-term follow-up. Lastly, while our study provided information on the treatment patterns for HK, it did not assess the efficacy, safety, and tolerability of the different treatments.

Conclusions

In conclusion, HK occurrence and recurrence are highly frequent among a real-world cohort of Chinese patients with ESRD, despite adequate HD. The high frequency of and short time to HK recurrence allude to the chronic nature of HK in these patients; hence, effective long-term control of sK on non-dialysis days is crucial. Yet, potassium binders to manage HK and limit sK fluctuations remain underused.

Funding Statement

The PRECEDE-K study was supported by AstraZeneca Investment China Co., Shanghai, China. This study was also funded by the National Natural Science Foundation of China (82070693), the Clinical Research Plan of Shanghai Hospital Development Center (No. SHDC2020CR3029B), and the Multicenter Clinical Research Project of Shanghai Jiao Tong University School of Medicine (DLY201805).

Authors’ contributions

ZN and HJ contributed equally as first authors. ZN conceived and designed the study. LZhang, LY, GS, LZuo, SQ, XZ, QZ, WY, QL, YR, HP, JX, QY, and QC acquired, analyzed, or interpreted the data. The PRECEDE-K study investigators conducted the statistical analysis. HJ and YS wrote and revised the manuscript; ZN critically revised the manuscript for important intellectual content; all authors critically reviewed the manuscript and approved the final version for submission. ZN was the principal investigator; HJ and RL were subprincipal investigators.

Disclosure statement

Yifan Shi is an employee of AstraZeneca Investment China Co, Shanghai, China. All the other authors declare that they have no competing interests.

Ethics statement

This study was conducted in accordance with the Declaration of Helsinki, the International Conference on Harmonization’s Guidelines for Good Clinical Practice, and applicable local legislation on non-interventional studies and/or observational studies. All patients provided written and oral informed consent forms pre-enrollment. The protocol and all its amendments were approved by the Shanghai Jiao Tong University School of Medicine Renji Hospital Ethics Committee (2020-040) and the ethics committee of each participating center.

Data availability statement

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.