Hyperkalemia in Pediatric Chronic Kidney Disease

Introduction

Potassium homeostasis is disrupted in chronic kidney disease (CKD) and can result in the development of hyperkalemia, a potentially life-threatening complication, through multiple mechanisms. This includes reduced potassium clearance, extracellular shifting due to CKD-associated metabolic acidosis, excessive dietary intake, and effects from the use of certain medications including renin-angiotensin-aldosterone system blockade (RAAS; angiotensin converting enzyme inhibitors-ACEi- and angiotensin receptor blockers-ARB), beta blockers, and diuretics [1,2]. Among adults with CKD, the prevalence of hyperkalemia is 14-20% and it confers significant morbidity including added medication burden, introduction of stringent dietary restrictions, increased hospitalizations and the possible initiation of kidney replacement therapy [3]. Moreover, hyperkalemia leads to an increased risk of CKD progression, major adverse cardiovascular events including arrhythmias, and higher mortality [4,5]. Frequently, hyperkalemia also leads to discontinuation of important therapies in patients with CKD, specifically ACEi and ARBs. Recent work in children has identified the importance of continuation of ACEi and ARB therapy whose use has been shown to slow the progression of CKD and delay time to kidney replacement therapy [6-9].

There is very limited published data regarding hyperkalemia in children with CKD. One single-center experience of 366 pediatric subjects reported the prevalence of hyperkalemia in Stage 4-5 CKD to be 18.8% [10]. However, larger scale studies assessing the prevalence of hyperkalemia in pediatric patients across all CKD stages have not previously been undertaken. Furthermore, there is no published literature on the relationship between hyperkalemia and CKD stage, demographics, laboratory parameters including proteinuria and acid-base status, and the use of potassium-influencing medications in children. Importantly, identification of risk factors for hyperkalemia in pediatric CKD may help clinicians anticipate the abnormality in a timely manner and pre-emptively initiate potassium-lowering therapy or institute dietary modifications prior to the development of a potentially life-threatening arrhythmia. Knowledge of the burden of hyperkalemia with respect to medication use may also help inform decisions about when to institute these changes to mitigate the hyperkalemic side effects of important disease modifying therapies.

In this cross-sectional study, we sought to characterize the serum potassium status in children with mild to moderate CKD using a large cohort of subjects followed longitudinally. Specifically, we aimed to investigate the prevalence of hyperkalemia with respect to patient factors including demographic information, CKD stage, CKD etiology, and laboratory evaluations. Risk factors for hyperkalemia were identified through an adjusted analysis to provide clinicians with specific populations of patients in whom focused monitoring and treatment could be instituted.

Methods

Study Population

Data were from the Chronic Kidney Disease in Children (CKiD) study, a prospective multicenter cohort study with participants from the US and Canada with study design details previously described [11]. Briefly, the CKiD study contains longitudinal data collected at annual visits from children enrolled in the study at ages 6 months to 16 years with an initial eGFR of 30 to 90mL/min/1.73m². Data collected with each visit includes demographics, clinical diagnoses, medical history, therapy use, physical examination, and laboratory values. The CKiD protocol and sites were approved by local site Institutional Review Boards (IRBs) and informed consent and assent were provided by all parents and appropriately-aged participants. For this study, 5183 study visits contributed by 1050 unique participants were included.

Variables and Analyses

Hyperkalemia was defined as a serum potassium level greater than or equal to 5.5mmol/L. Estimated glomerular filtration rate (eGFR) was calculated using the U25 equation [12]. CKD Stages were defined by eGFR based on the 2002 KDIGO classification guidelines and categorized into CKD Stage 1/2 (eGFR ≥60mL/min/1.73m²), Stage 3a (eGFR 45-59mL/min/1.73m²), Stage 3b (eGFR 30-44mL/min/1.73m2), and Stages 4/5 (eGFR <30mL/min.1.73m²) [13]. CKD etiology was defined as glomerular (G) (including hereditary nephritis, focal segmental glomerulosclerosis, hemolytic uremic syndrome, or other glomerular disease) or non-glomerular (NG) (including obstructive/reflux nephropathy, hypoplasia, dysplasia, cystic kidney disease, interstitial nephritis, and other congenital anomalies of the kidney and urinary tract-CAKUT). Serum CO2 level was used as a surrogate for acid-base status and was characterized as low CO2 (<22mmol/L), normal CO2 (≥ 22 to < 26mmol/L) and high CO2 (≥26mmol/L). Proteinuria was defined by the urine protein to creatinine ratio (UPCR) and categorized as none (<0.2g/g), non-nephrotic range (0.2 to 1.9g/g) and nephrotic-range (≥ 2.0g/g).

Sociodemographic and clinical characteristics were presented as a percentage with count (n) for categorical variables and mean with standard deviation (SD) or median with interquartile ranges [IQR] for continuous variables. Differences in variables by glomerular or non-glomerular CKD etiologies were assessed using Chi-square tests for categorical variables and T-tests or Kruskal-Wallis tests for continuous variables. The proportion of participants with hyperkalemia within a specific group and associated 95% confidence intervals (CI) were estimated by fitting repeated measures logistic regressions with group indicators as independent variables and converting the log odds of hyperkalemia to proportions. For example, three dummy indicators for CKD stages were used in the logistic regression model. Generalized estimating equations (GEE) were used to handle within-person correlations for appropriate standard error estimation.

A series of univariate and multivariate models with GEE were used to examine associations between the presence of hyperkalemia and putative risk factors chosen a priori based on adult data and pathophysiology including age, sex, self or parent-reported ethnicity (Hispanic and African American), low socioeconomic status (defined by the presence of public insurance, annual household income of <$36,000, or maternal education less than college), CKD diagnosis, CO2-defined acid-base status, CKD stage, UPCR-defined proteinuria level, and medication use including alkali therapy, beta blockers, ACEi/ARBs, diuretics, and potassium binders. Variables with a p value of <0.1 in the univariate models were then included in the multivariate model. Statistical significance was assessed at p <0.05.

A post-hoc subanalysis of participants with advanced CKD (Stage 4/5) was conducted to evaluate the association between self-reported ACEi/ARB use, with and without diuretic therapy, and hyperkalemia. ACEi/ARB alone, diuretic alone, and combined ACEi/ARB plus diuretic therapy category indicators were used in the regression model and adjusted for the aforementioned factors. P values were obtained from type 3 tests for the overall difference in hyperkalemia across all groups. All analyses were conducted with SAS 9.4 (SAS Institute, Cary NC).

Results

Table 1 presents demographic and clinical characteristics of the participants. The mean age for subjects was 13.1 years, 62.7% were male, 19.4% were African American, and 13.5% were Hispanic. 76.6% of person-visits were from subjects with non-glomerular etiologies of CKD. CKD stages were distributed as 28.8% with Stage 1/2, 26.1% with Stage 3a, 26.4% with Stage 3b, and 18.7% with Stage 4/5. The median eGFR was 48mL/min/1.73m² [IQR 34-63 mL/min/1.73m²] and 25.8% had low CO2 levels. Self-reported medication use within this cohort revealed that 54.2% of subjects received ACEi or ARB therapy, whereas only 1.3% reported treatment with a potassium binder.

Table 1.

Demographic, clinical and laboratory characteristics of participants.

Characteristics	% (person-visits) n = 5183
Age, years, mean (SD)	13.1 (5.0)
Male sex	62.7 (3251)
African-American ethnicity	19.4 (1007)
Hispanic ethnicity	13.5 (697)
Income <$36,000/year	20.1 (1039)
Maternal education < college	57.3 (2961)
Public Insurance	24.1 (1248)
CKD Etiology
Glomerular	23.3 (1209)
Non-glomerular	76.6 (3974)
UPCR, median [IQR]	0.3 [0.1, 1.0]
Proteinuria Level (g/g)
No proteinuria (<0.2)	36.4 (1886)
Non-nephrotic (0.3-1.9)	51.7 (2680)
Nephrotic (>2.0)	11.9 (617)
U25eGFR, median [IQR]	48 [34, 63]
CKD Stage (eGFR, mL/min/1.73m2)
1/2 (≥60)	28.8 (1492)
3a (45-50)	26.1 (1353)
3b (30-44)	26.4 (1368)
4/5 (<30)	18.7 (970)
Potassium, mmol/L, median [IQR]	4.3 [4.1, 4.7]
Hyperkalemia (K >5.5mmol/L)	2.2 (116)
Acid-Base Status (CO2, mmol/L)
High (CO2 ≥ 26)	27.9 (1445)
Normal (CO2 ≥22 to <26)	46.3 (2399)
Low (CO2 <22)	25.8 (1339)
Self-Reported Therapy Use
Alkali	26.4 (1369)
ACEi/ARB	54.2 (2809)
Beta blocker	5.6 (292)
Diuretic	5.8 (300)
Potassium binder	1.3 (67)

Median serum potassium levels varied by CKD stage. Median serum potassium levels were significantly higher in subjects with advanced stage CKD (Stage 4/5, 4.5mmol/L, IQR 4.1-5.0mmol/L), and the percentage of person-visits with hyperkalemia was also higher in subjects with CKD Stage 4/5 (6.6%, 95% CI 4.8%-8.2%) (Table 2) compared to those with less advanced CKD stages. When evaluating hyperkalemia with respect to CKD etiology, these trends remained with higher median serum potassium levels and a greater percentage of person-visits with hyperkalemia in Stage CKD 4/5 when compared to milder stages of CKD, irrespective of glomerular or non-glomerular CKD etiology (Table 2). However, 14.3 percent of person-visits with Stage 4/5 CKD of glomerular etiology had hyperkalemia (95% CI 9.4%-21.1%) with a median serum potassium level of 4.9mmol/L (IQR 4.3-5.3mmol/L), whereas 5.1% of person-visits with Stage 4/5 CKD of non-glomerular etiology had hyperkalemia (95% CI 3.5%-6.8%) with a median serum potassium level of 4.5mmol/L (IQR 4.1-4.9 mmol/L) (Table 2). There was no difference in the median serum potassium levels or percentage of visits with hyperkalemia between male and female subjects (data not shown). When varied by race and ethnicity, Hispanic participants had a slightly higher percentage of visits with hyperkalemia when compared to non-Hispanic participants (3.6% vs. 2.0%, p=0.023, data not shown).

Table 2.

Median serum potassium levels and percentage of assessments with hyperkalemia by CKD stage and etiology, n = 5183

Stage (eGFR, mL/min/1.73m2)	1/2 (60+)	3a (45-59)	3b (30-44)	4/5 (<30)	p-value
Overall, N (# of visits)	1492	1353	1368	970
Potassium, mmol/L [IQR]	4.2 [4.0, 4.5]	4.3 [4.1, 4.6]	4.4 [4.1, 4.7]	4.5 [4.1, 5.0]	<0.001
Hyperkalemia, % (n) [95% CI]	0.6 (9) [0.3, 1.2]	0.8 (11) [0.5, 1.5]	2.3 (32) [1.6, 3.4]	6.6 (64) [4.8, 8.2]	<0.001
Glomerular, n *	538	290	227	154
Potassium, mmol/L [IQR]	4.3 [4.0, 4.5]	4.4 [4.1, 4.8]	4.5 [4.2, 4.9]	4.9 [4.3, 5.3]	<0.001
Hyperkalemia, % (n) [95% CI]	1.1 (6) [0.5, 2.5]	1.0 (3) [0.3, 3.1]	2.2 (5) [0.9, 5.1]	14.3 (22) [9.4, 21.1]	0.002
Non-glomerular, n *	954	1063	1141	816
Potassium, mmol/L [IQR]	4.2 [4.0, 4.5]	4.3 [4.1, 4.6]	4.4 [4.1, 4.7]	4.5 [4.1, 4.9]	<0.001
Hyperkalemia, % (n) [95% CI]	0.3 (3) [0.1, 1.0]	0.8 (8) [0.4, 1.5]	2.4 (27) [1.6, 3.5]	5.1 (42) [3.5, 6.8]	<0.001

^*Glomerular or nonglomerular etiology of CKD, # of study visits

When stratifying person-visits by CO2-defined acid-base status, median serum potassium levels were highest and a greater percentage of person-visits with hyperkalemia occurred in the low CO2 category (median serum potassium level 4.5mmol/L, IQR 4.1-4.9mmol/L and 4.8% of person-visits with hyperkalemia) when compared to visits with higher CO2 levels (Table 3). Subjects with nephrotic-range proteinuria also had higher median serum potassium levels (4.4mmol/L, IQR 4.0-4.8mmol/L) and a higher percentage of person-visits with hyperkalemia (3.3%) compared to those person-visits with lesser amounts or no proteinuria (Table 3).

Table 3.

Median serum potassium levels and percentage of assessments with hyperkalemia by CO2 level and level of proteinuria.

CO2 mmol/L	High >26	Normal 22-26	Low <22	p-value
N (# of visits)	1407	2333	1304
Potassium, mmol/L [IQR]	4.3 [4.0, 4.5]	4.3 [4.1, 4.6]	4.5 [4.1, 4.9]	<0.001
Hyperkalemia, % (n)	0.4 (5)	1.9 (45)	4.8 (63)	<0.001
Level of Proteinuria UPCR, g/g	None <0.2	Non-nephrotic 0.2–2.0	Nephrotic >2.0	p-value
N (# of visits)	1407	2611	608
Potassium, mmol/L [IQR]	4.3 [4.1, 4.6]	4.4 [4.1, 4.7]	4.4 [4.0, 4.8]	<0.001
Hyperkalemia, % (n)	1.3 (24)	2.6 (69)	3.3 (20)	0.002

Univariate analysis identified several clinical features associated with hyperkalemia including Hispanic ethnicity, low socioeconomic status, CKD etiology, low acid-base status (CO2-based), advanced CKD stage, proteinuria level, and self-reported therapy with alkali, beta blocker, and ACEi/ARB use (Table 4). Potassium binder and diuretic therapy were also associated with hyperkalemia (Table 4). In the multivariate model adjusting for these variables, CKD stage was significantly associated with an increased risk of hyperkalemia; specifically, compared to those person-visits with CKD Stage 1/2, CKD Stage 3b was associated with 3.25-fold increased odds of hyperkalemia (OR 3.25, 95% CI 1.48-7.15, p=0.003) and CKD Stage 4/5 was associated with a 9-fold greater odds (OR 9.17, 95% CI 4.02-20.89, p<0.001). CO2-based acid-base status was also independently associated with hyperkalemia. Among subject visits with CO2 less than 22mmol/L, the odds of having hyperkalemia was 7.72 times higher than in those with CO2 values of 26mmol/L or higher (OR 7.72, 95% CI 3.05-19.54, p<0.001). Interestingly, participants with a normal CO2 level also had a higher odds of hyperkalemia when compared to those with elevated CO2 levels (OR 4.62, 95% CI 1.94-11.06, p=0.001). In contrast, a non-glomerular CKD etiology was associated with a nearly 50% decreased odds of developing hyperkalemia (OR 0.52, 95% CI 0.34-0.80, p=0.003). Race and ethnicity, low socioeconomic status, and proteinuria level as defined by UPCR were not significant risk factors for hyperkalemia.

Table 4.

Risk factors associated with the presence of hyperkalemia.

Characteristics	Univariate model		Multivariate model
	Odds ratio (95% CI)	p-value	Odds ratio (95% CI)	p-value
Age, years	1.0 (0.96, 1.03)	0.799
Male	1.13 (0.72, 1.79)	0.593
Race/Ethnicity
African American	0.76 (0.43, 1.32)	0.328
Hispanic	1.80 (1.09, 2.97)	0.023	1.12 (0.68, 1.84)	0.669
Low socioeconomic status	1.73 (1.08, 2.76)	0.023	1.51 (0.94, 2.43)	0.091
Non-glomerular CKD†	0.67 (0.43, 1.04)	0.073	0.52 (0.34, 0.80)	0.003
Acid-Base Status ‡		<.0001*		<0.0001*
Normal	5.63 (2.31, 13.72)	<.001	4.63 (1.94, 11.06)	0.001
Low	14.69 (5.85, 36.93)	<.0001	7.72 (3.05, 19.54)	<0.0001
CKD Stage §		<.0001*		<0.0001*
3a	1.35 (0.56, 3.25)	0.502	1.26 (0.51, 3.10)	0.615
3b	3.95 (1.87, 8.35)	<.001	3.25 (1.48, 7.15)	0.003
4/5	11.64 (5.74, 23.61)	<.0001	9.17 (4.02, 20.89)	<0.0001
Proteinuria Level ∥		0.003*		0.117*
Non-nephrotic	2.03 (1.20, 3.43)	0.009	0.94 (0.53, 1.66)	0.832
Nephrotic	2.49 (1.29, 4.83)	0.007	0.54 (0.25, 1.16)	0.114
Self-Reported Therapy
Alkali	2.31 (1.56, 3.43)	<.0001	1.12 (0.72, 1.73)	0.628
Beta blocker	1.97 (1.04, 3.73)	0.036	1.24 (0.64, 2.38)	0.524
ACEi/ARB	2.26 (1.45, 3.51)	<.001	2.14 (1.36, 3.37)	0.001
Diuretic	2.29 (1.18, 4.46)	0.014	1.32 (0.73, 2.40)	0.362
Potassium binder	5.36 (2.09, 13.73)	0.001	3.76 (1.52, 9.27)	0.004

^*p-value from type 3 test for the overall difference in hyperkalemia across all groups

^†Reference: Glomerular CKD

^‡.Reference: High, CO2 ≥ 26mmol/L; Normal CO2 ≥22 to <26mmol/L, low CO2 <22mmol/L)

^§Reference: Stage 1/2, eGFR ≥60mL/min/1.73m²

^∥Reference: no proteinuria, UPCR < 0.2g/g

Of the self-reported medications included in the multivariate model, subjects receiving ACEi/ARB therapy had a significantly greater risk of hyperkalemia (OR 2.14, 95% CI 1.36-3.37, p=0.001) (Table 4). Use of alkali therapy, beta blockers, and diuretics was not associated with a higher risk of hyperkalemia. Not surprisingly, there was a significant independent relationship between hyperkalemia and the use of potassium binder therapy (OR 3.75, 95% CI 1.52-9.27, p=0.004).

Finally, in the subanalysis of 970 person-visits with CKD 4/5, 41.7% (n=404) reported use of ACEi/ARB alone, 5.6% (n=54) reported diuretic use alone, 6.3% (n=61) reported ACEi/ARB and diuretic use, and 46.5% (n=451) did not self-report use of either ACEi/ARB or diuretic therapy. In these subjects, multivariate-adjusted analysis revealed that compared to those without ACEi/ARB or diuretic use, those with ACEi/ARB use had a significantly higher risk of hyperkalemia (OR 2.48, 95% CI 1.28-4.80, p=0.007). Combined ACEi/ARB and diuretic use also had an increased risk of hyperkalemia (OR 3.60, 95% CI 1.36-9.48, p=0.010). Diuretic use alone was not associated with a higher risk of hyperkalemia (OR 1.62, 95% CI 0.44-5.97, p=0.478).

Discussion

Hyperkalemia frequently complicates CKD in adults due to the kidney’s crucial role in potassium homeostasis and its development is associated with increased morbidity and mortality. Some studies in adult CKD patients have provided evidence of its detection in almost a quarter of patients [3]. Whereas the prevalence of hyperkalemia in mild to moderate pediatric CKD has previously been uncharacterized, analysis of CKiD data has provided the opportunity to describe the status of potassium in pediatric CKD in relation to multiple patient, laboratory and therapy variables and to identify potential risk factors for the development of hyperkalemia.

The results of this study demonstrate several important trends for hyperkalemia in pediatric CKD. In the unadjusted analyses, more advanced stages of CKD were associated with higher median serum potassium levels and a higher percentage of visits with hyperkalemia. In particular, hyperkalemia was detected in 6.6 percent of person visits from subjects with Stage 4 and 5 CKD compared to only 0.6 to 2.3 percent of person visits from subjects with Stage 1 to Stage 3b CKD. The multivariate analysis further revealed that advanced CKD stage (Stage 4/5) alone confers over a nine-fold increased risk of hyperkalemia. This finding aligns with results from a large adult study demonstrating a 25% increased risk of hyperkalemia for every 5mL/min/1.73m² decrease in eGFR [5]. In turn, these data highlight patients at greatest risk for hyperkalemia whose management should include close monitoring and prompt therapeutic intervention to decrease the risk of adverse outcomes.

In a prior pediatric study, Wong et al noted an almost 20% prevalence of hyperkalemia in a small cohort of participants with Stage 4 and 5 CKD [10]. A similar percentage of visits with hyperkalemia was seen in our cohort of participants with advanced CKD secondary to a glomerular disorder. Our data revealed that in pediatric patients with a glomerular diagnosis and Stage 4/5 CKD, hyperkalemia was present in 14.3% of assessments, significantly more frequently than what was experienced by participants with NG disease. The adjusted analysis further confirmed that those with a glomerular etiology of CKD had a significantly higher odds than those with NG disease. While the pathophysiology behind this is not entirely clear, the substantial urine output that is commonly associated with CAKUT and other non-glomerular etiologies of CKD may lead to enhanced potassium excretion.

A low CO2 level as a surrogate for acidosis also significantly increased the risk of hyperkalemia in our pediatric CKD subjects, which is to be expected given the physiologic effects of acidosis on extracellular potassium movement. Acidosis has previously been shown to negatively impact growth and CKD progression in children and its relationship with potassium status is yet another reason why this modifiable risk factor should be aggressively treated [14-16]. Interestingly, subjects with normal CO2 levels also had a more than four times greater risk of hyperkalemia when compared to subjects with CO2 levels above 26mmol/L. Whereas the mechanism behind this finding was not explained by the available data, higher CO2 levels in patients with CKD may confer some additional protection against hyperkalemia based on the previously noted physiology of potassium movement [6]. While dietary intake of potassium may contribute to the development of hyperkalemia as well, neither comprehensive nutrient data nor specific information pertaining to nutrition counseling has routinely been collected from the CKiD cohort, precluding assessment of their roles in our analysis.

We did find a small but statistically significant higher proportion of person-visits with hyperkalemia among participants of Hispanic ethnicity. A study by Kim et al showed a similar trend with higher serum potassium levels in Hispanic patients on hemodialysis when compared to Black or White individuals [17]. Interestingly, these subjects did not have a higher burden of mortality, leading investigators to postulate that Hispanic patients may have a higher “set point” in their range of potassium tolerance [17,18].

Noteworthy was our finding that concurrent ACEi or ARB therapy was associated with a significantly increased risk of hyperkalemia when compared to those not receiving this therapy, highlighting another important group of patients who should be monitored closely and/or started on potassium-lowering therapies or possibly a modified dietary potassium intake. Subgroup analysis shows that in those with CKD 4/5, concomitant therapy with a diuretic was not associated with a lower risk of hyperkalemia. The benefits associated with continuation of RAAS inhibition therapy in patients with CKD are numerous and serve as the topic of many recent CKD-related studies. A retrospective study of almost 80,000 Canadian adults with CKD and hyperkalemia showed that RAAS inhibitor discontinuation was associated with higher mortality and more frequent cardiovascular events, suggesting that efforts should be made to maintain patients on these medications despite the increased risk of potassium abnormalities, if possible [19]. Very little information regarding the use of these therapies in children and their safety and efficacy in patients with Stage 4/5 CKD is available. However, in a recent observational study in pediatric CKD, 23% of participants discontinued RAAS inhibitor therapy due to hyperkalemia [8]. Though potassium levels decreased following cessation of RAAS inhibition, eGFR was observed to decline more rapidly in those who discontinued therapy when compared to the rate of eGFR decline experienced by those participants who did not discontinue therapy [8]. These findings are a call to action for the pediatric nephrology community to engage in prospective studies and generate safety and efficacy evidence upon which strategies to reduce hyperkalemia while maintaining RAAS inhibitor therapy can be well defined and made available to children. Current approaches include dietary modification, potassium exchange resin therapy (specifically sodium polystyrene sulfonate), and initiation of non-potassium-sparing diuretics. Medications which have already proven beneficial in adults and may ultimately prove to be safe and beneficial in children as well include the mineralocorticoid receptor antagonist finerenone and the potassium binding resins patiromer and zirconium cyclosilicate, all of which are currently not labeled for pediatric use, but are undergoing evaluation in the pediatric CKD population [20].

Of interest in the current study, the use of potassium binder therapy was associated with hyperkalemia when controlling for CKD stage, CKD etiology and other medication use. This finding is likely attributable to the high-risk nature of the population for whom the therapy is prescribed. Reverse causation may be present with clinicians recognizing the risk for hyperkalemia in this group and initiating treatment in those felt to be at the highest risk. However, this may also indicate at least some treatment resistance as hyperkalemia was ascertained after potassium binder therapy use was self-reported.

To our knowledge, this is the largest study of potassium status in pediatric patients with CKD undertaken to date. A strength of this study is the inclusion of a large number of well described subjects with both glomerular and non-glomerular kidney disorders followed longitudinally with adequate representation across all stages of CKD. We also acknowledge that there are limitations to this study. Self-reported diuretic use in this study encompassed loop, thiazide, and potassium-sparing diuretics, whose differential effects on potassium handling may have confounded the overall relationship between this category of medication and hyperkalemia making it difficult to draw conclusions regarding its use.

Furthermore, while we are unable to comment on the contribution of the dietary intake of potassium or nutritional counseling to our results, the importance of dietary management to achieve normal serum potassium levels in pediatric CKD cannot be overstated. The topic of dietary management for hyperkalemia was addressed in the National Kidney Foundation’s Kidney Disease Outcomes Quality Initiative (KDOQI) nutrition guidelines for children with CKD more than a decade ago [21]. A recent guideline by the Pediatric Renal Nutrition Taskforce provides further clinical practice recommendations for the management of potassium in children with CKD [22].

In conclusion, hyperkalemia is a serious complication of pediatric CKD, and children with advanced stages of CKD and glomerular disease, low serum CO2 level, or using ACEi or ARB therapy are at particular risk. These data should help clinicians focus on high-risk patient groups and should serve as an impetus for further research into potassium-lowering therapies for pediatric patients with CKD.