L-type calcium channel blocker use and proteinuria among children with chronic kidney diseases

Abstract

Background

Hypertension is common among children with chronic kidney disease (CKD), and dihydropyridine calcium channel blockers (dhCCBs) are frequently used as treatment. The impact of dhCCBs on proteinuria in children with CKD is unclear.

Methods

Data from 722 participants in the Chronic Kidney Disease in Children (CKiD) longitudinal cohort with a median age of 12 years were used to assess the association between dhCCBs and log transformed urine protein/creatinine levels as well as blood pressure control measured at annual visits. Angiotensin-converting enzyme inhibitor (ACEi) and angiotensin receptor blocker (ARB) use was evaluated as an effect measure modifier.

Results

Individuals using dhCCBs had 18.8% higher urine protein/creatinine levels compared to those with no history of dhCCB or ACEi and ARB use. Among individuals using ACEi and ARB therapy concomitantly, dhCCB use was not associated with an increase in proteinuria. Those using dhCCBs had higher systolic and diastolic blood pressures.

Conclusions

Use of dhCCBs in children with CKD and hypertension is associated with higher levels of proteinuria and was not found to be associated with improved blood pressure control.

Introduction

Hypertension affects approximately half of children with chronic kidney disease (CKD) and its prevalence increases as CKD progresses towards CKD stage 5 (CKD 5) [1]. In addition, hypertension is a major risk factor for development of cardiovascular morbidity as well as deterioration in kidney function over time in these children [2–4]. Despite this, data from the Chronic Kidney Disease in Children (CKiD) cohort suggest that blood pressure remains inadequately controlled in this patient population [5, 6]. Although a variety of antihypertensive medications are available, L-type dihydropyridine calcium channel blockers (dhCCBs), such as amlodipine, isradipine, and nifedipine, are commonly used for blood pressure control in children with CKD [1].

There is strong evidence that pharmacologic treatment of hypertension with angiotensin-converting enzyme inhibitors (ACEi) and angiotensin receptor blockers (ARB) can slow progression of CKD as well as lead to normalization of left ventricular function and left ventricular mass [4, 7]. ACEi and ARB also assist with control of proteinuria [8]. However, studies have shown that despite the potential benefits of not doing so, many clinicians discontinue ACEi/ARB use as kidney function declines due to concern for worsening glomerular filtration rate (GFR) and/or hyperkalemia [9–11]. Due to this concern, dhCCBs may be increasingly used instead. Additionally, for individuals with difficult to treat hypertension, dhCCBs may be used in combination with ACEi/ARB therapy.

Studies conducted in adults demonstrate that L-type dhCCBs may actually be associated with accelerated progression of kidney disease due to an increase in pre-existent proteinuria. In adults with diabetic nephropathy, amlodipine was associated with a 6% increase in proteinuria compared to baseline, as well as a 23% higher incidence of doubling of serum creatinine, development of CKD 5 or death [12]. The African American Study of Kidney Disease and Hypertension (AASK) demonstrated a 58% increase in the geometric mean of urine protein to creatinine ratio in those receiving amlodipine [13]. Other studies have failed to demonstrate a statistically significant change in proteinuria related to dhCCB use but have demonstrated a plateau or slight increase in proteinuria with amlodipine compared to other antihypertensive treatments [14, 15]. There are no current data available in children regarding any effect of dhCCB use on proteinuria or CKD progression. The purpose of this analysis was to evaluate how dhCCB use relates to proteinuria, as well as blood pressure control, among children with CKD, while appropriately accounting for ACEi/ARB use. We hypothesized that individuals with hypertension who receive dhCCBs would have increased proteinuria compared to those not receiving dhCCBs.

Methods

Study population

The CKiD study is a longitudinal observational cohort study that enrolled children with a diagnosis of CKD between the ages of 1 and 16 and a GFR < 90 ml/min/1.73 m² from sites in the US and Canada. Participants attended annual visits and follow a structured protocol to collect data on medical history, current health, medication use, and clinical variables, including biological samples for kidney health and function (e.g., urine and blood samples) as well as blood pressure measurements. Self- or parental-reported medication use was collected at each visit. A detailed description of the study design has been previously published [16].

For this analysis, the sample population comprised all person-visits in which age was younger than 18 years, with complete data on dhCCB use (exposure), urine protein and blood pressure (outcomes), ACEi/ARB use (effect measure modifiers), and eGFR, as a marker of disease severity.

Exposure and effect measure modifiers

The primary exposure was self-reported use of dhCCBs (including amlodipine, isradipine, and nifedipine) to estimate the effect of this therapy on urine protein levels and blood pressure. This exposure was investigated as those who, across two sequential annual visits, never used, used, or discontinued dhCCB therapy. The two annual visits were defined as the index, or current visit, and the previous visit (approximately one year earlier). Since ACEi/ARBs are used to treat hypertension and proteinuria, and dhCCBs are primarily used as an antihypertensive therapy only with a hypothesized adverse effect on proteinuria, we investigated ACEi/ARB use as a potential modifier of the effect of dhCCBs. Specifically, we investigated proteinuria as a continuous variable, using the following categories: no dhCCB use, current dhCCB, and discontinued dhCCB among those who, across the same two annual visits, did not use ACEi/ARBs, used ACEi/ARBs, or discontinued ACEi/ARBs.

Outcome

Log-transformed urine protein to creatinine ratio (UPC) was the primary outcome of interest. Urine protein to creatinine ratio was used as a continuous variable ranging from no proteinuria to nephrotic range proteinuria. Secondary outcomes included blood pressure, which was calculated as the average of three auscultatory measurements at least 30 seconds apart, obtained using an aneroid sphygmomanometer. Details of the CKiD blood pressure measurement protocol have previously been described [1]. Blood pressure values were transformed to systolic blood pressure (SBP) z-scores and diastolic (DBP) z-scores adjusting for age, gender, and height according to the 2017 American Academy of Pediatrics guidelines [17]. SBP and DBP z-scores were then used as the outcome to describe the association between CCB use and blood pressure control.

Covariates

We considered several confounders of the hypothesized relationship between dhCCB and proteinuria. Disease severity at the previous visit was considered a strong confounder and was quantified using eGFR (in the log scale and centered at 50 ml/min/1.73 m²) and proteinuria (in the log scale and centered at 0.5 mg/mg Cr). Second, high blood pressure is an indication for dhCCB therapy and is known to be strongly associated with proteinuria [18]; therefore, previous year SBP and DBP z-scores were included as continuous variables. In total, for this interaction between two variables with three response types (i.e., for dhCCB and ACEi/ARB, the response types were no use, current use, and discontinued use), there were nine groups, with the reference group being participants who did not use either dhCCBs or ACEi/ARB at the previous visit or the current annual visit. This single reference group allowed for estimation of association of dhCCB or ACEi/ARB use alone, as well as all combinations of ACEi/ARB and dhCCB use/discontinuation.

Lastly, age as a continuous variable (centered at 13 years), sex, and race (African American vs. non-African American) were also included as confounders. Age was considered a surrogate for disease duration and self-management of disease. There was imbalance in the distribution of sex in which there were proportionally fewer boys who discontinued dhCCB use compared to those who used dhCCBs, and African American participants were more likely to use dhCCBs than non-African Americans. Previous CKiD work has shown that African American race was associated with earlier time to CKD 5 [19]. We did not adjust for time-varying variables (GFR, proteinuria, and SBP and DBP z-scores) at the time of reported dhCCB use and proteinuria measurement, but rather we adjusted for these variables at the previous annual visit (i.e., lagged) to ensure that we did not control for biomarkers that were measured after the exposure (dhCCB use or ACEi/ARB use) occurred.

Statistical analyses

The unit of analysis was pairs of person-visits. Since several variables were included from the previous annual visit (dhCCB use, ACEi/ARB use, proteinuria, GFR, and SBP and DBP z-scores), the analytic study entry technically occurred at the second annual visit. A linear model was fit, using the ‘geeglm’ function from the ‘geepack’ package in R, with log-transformed UPC as the dependent variable and generalized estimating equations (GEE) to account for the dependencies of within-person repeated measurements. The model was of the form [20]:

$$\begin{gathered} \log \left(UPC\right)\mathrm{=}{\beta }_0+{\beta }_1\times \mathrm{C}\mathrm{C}{\mathrm{B}}_{.1}\mathrm{A}\mathrm{C}\mathrm{E}\mathrm{i}\mathrm{A}\mathrm{R}{\mathrm{B}}_{00}+{\beta }_2\times \mathrm{C}\mathrm{C}{\mathrm{B}}_{10}\mathrm{A}\mathrm{C}\mathrm{E}\mathrm{i}\mathrm{A}\mathrm{R}{\mathrm{B}}_{00} \\ +{\beta }_3\times \mathrm{C}\mathrm{C}{\mathrm{B}}_{00}\mathrm{A}\mathrm{C}\mathrm{E}\mathrm{i}\mathrm{A}\mathrm{R}{\mathrm{B}}_{.1}+{\beta }_4\times \mathrm{C}\mathrm{C}{\mathrm{B}}_{.1}\mathrm{A}\mathrm{C}\mathrm{E}\mathrm{i}\mathrm{A}\mathrm{R}{\mathrm{B}}_{.1}+{\beta }_5\times \mathrm{C}\mathrm{C}{\mathrm{B}}_{10}\mathrm{A}\mathrm{C}\mathrm{E}\mathrm{i}\mathrm{A}\mathrm{R}{\mathrm{B}}_{.1} \\ +{\beta }_6\times \mathrm{C}\mathrm{C}{\mathrm{B}}_{00}\mathrm{A}\mathrm{C}\mathrm{E}\mathrm{i}\mathrm{A}\mathrm{R}{\mathrm{B}}_{10}+{\beta }_7\times \mathrm{C}\mathrm{C}{\mathrm{B}}_{.1}\mathrm{A}\mathrm{CEi}\mathrm{A}\mathrm{R}{\mathrm{B}}_{10}+{\beta }_8\times \mathrm{C}\mathrm{C}{\mathrm{B}}_{10}\mathrm{A}\mathrm{CEi}\mathrm{A}\mathrm{R}{\mathrm{B}}_{10}+{\beta }_{\mathrm{Z}}\times \mathit{Z} \end{gathered}$$

where subscripts xy (i.e., CCB_xy and ACEiARB_xy) indicate patterns of use between consecutive visits. Subscript “x” represents medication use at the previous visit, and “y” represents medication use at the current visit. Medication use is indicated by “1,” no use is indicated by “0,” and “•” indicates that medication may or may not have been used. Thus, a subscript “00” indicates no previous or current medication use; “•1” indicates current use, regardless of previous use; “10” indicates previous use, but not current use (i.e., discontinued use). Additional covariates Z in the model included the previous visit’s eGFR, UPC, SBP z-score, and DBP z-score, as well as age, sex, race, and underlying kidney disease diagnosis.

The primary exposure was current and previous use of dhCCBs with current and previous ACEi/ARB use as an effect measure modifier. This approach used longitudinal data to adjust for disease severity and therapy use at the previous year’s visit. Since ACEi/ARB discontinuation is common among those with lower GFR, and dhCCB use may be used as an alternative antihypertensive therapy, we further estimated differences for those who discontinued ACEi/ARB. In total, for this interaction between two variables with three response types (i.e., for dhCCB and ACEi/ARB, the response types were no use, current use, and discontinued use), there were nine groups, with the reference group being participants who did not use either dhCCBs or ACEi/ARB at the previous visit or the current annual visit. There were two groups with notably small sample sizes (i.e., < 20 person-visits): those with no previous or current ACEi/ARB use and who discontinued dhCCBs (n = 16 person-visits), and those who discontinued both ACEi/ARB and dhCCB use (n = 3 person-visits). The former group was indicated by gray shading in the figure to denote small sample size, and the latter group was not included in the analysis.

Statistical significance was assessed as p < 0.05. All analyses and graphics were conducted in R 3.6.1 (Vienna, Austria).

Results

A total of 722 CKiD participants provided 2503 person-visits to this analysis. Overall, at the first, or index, visit of the analysis, the median age was 12.0 years, 62% were boys and 28% had an underlying glomerular form of CKD. Table 1 describes the demographic and clinical person-visit characteristics at the index visit, stratified dhCCB use (no use, use, and discontinued). The majority of person-visits comprised participants with no previous or current dhCCB use (n = 2047, 81.8%). Those who used dhCCBs and discontinued dhCCB use comprised 16.1% and 2.1% of all person-visits, respectively. dhCCB users were more likely to be male, have lower eGFR, and a greater incidence of nephrotic range proteinuria. Participants who did not use dhCCBs had lower SBP and DBP percentiles and fewer were of African American race, compared to those who used dhCCBs and discontinued dhCCB use. Additionally, dhCCB users were less likely to use ACEi/ARBs and more likely to discontinue previous ACEi/ARB use, compared to those who never used or discontinued dhCCBs. A description of the distribution of person-visits by dhCCB and ACEi/ARB use is provided in Supplemental Figure 1. Supplemental Table 1 presents the descriptive statistics at the baseline visit for all participants, stratified by dhCCB use (no use vs. use).

Table 1.

Descriptive characteristics of study population across all person-visits (N = 2503 from 722 participants) stratified by dihydropyridine calcium channel blocker (dhCCB) use. n (%) or mean (SD) or median (interquartile range) for non-normally distributed variables. p-values are based on linear or logistic regression models for overall differences between the three exposure groups using generalized estimating equations to account for the dependencies of within-person repeated measurements.

	No previous or current dhCCB use (n = 2047 from 623 participants)	dhCCB use (n = 403 from 162 participants)	Discontinued dhCCB use (n = 53 from 51 participants)	p-value
Demographic characteristics
Age, years	12.5 (3.7)	12.3 (3.6)	12.6 (3.6)	0.729
Male	1270 (62.0)	267 (66.3)	27 (50.9)	0.108
Black race	316 (15.4)	113 (28.0)	17(32.1)	0.001
Kidney disease characteristics
Glomerular diagnosis	418 (20.4)	94 (23.3)	15 (28.3)	0.383
Duration of kidney disease, years	10.8 (4.3)	10.6 (4.3)	9.9 (4.7)	0.410
eGFRa, ml/min/1.73 m2	50 [37, 64]	39 [27, 59]	50 [33, 59]	<0.001
UPCa, mg/mgCr	0.3 [0.1, 0.9]	0.5 [0.2, 1.6]	0.4 [0.1, 1.0]	0.004
Nephrotic proteinuria (UPC > 2 mg/mgCr)	190 (9.3)	81 (20.1)	5 (9.4)	<0.001
Blood pressure
Previous diagnosis of hypertension	823 (40.2)	304 (75.4)	42 (79.2)	<0.001
SBP percentileb	52.2 (29.9)	73.1 (25.6)	65.3 (31.3)	<0.001
DBP percentileb	56.4 (27.6)	67.1 (28.5)	66.0 (26.2)	<0.001
Hypertension category
Normal BP	1545 (75.6)	190 (47.1)	34 (64.2)	<0.001
Elevated BP	195 (9.5)	65 (16.1)	6 (11.3)	<0.001
Stage 1 HTN	269 (13.2)	113 (28.0)	9 (17.0)	<0.001
Stage 2 HTN	36(1.8)	35 (8.7)	4 (7.5)	<0.001
Medication use
No previous or current ACEi/ARB use	690 (33.7)	144 (35.7)	15 (28.3)	0.606
ACEi/ARB Use	1272 (62.1)	210 (52.1)	35 (66.0)	0.054
Discontinued ACEi/ARB use	85 (4.2)	49 (12.2)	3 (5.7)	<0.001

^aMedian and interquartile range; p-values are based on log-transformed outcomes

^bp-values based on blood pressure z-scores converted from BP percentiles to achieve normally distributed outcomes

Differences in UPC by current and previous hypertension treatment, adjusted for covariates of interest at the previous visit, are displayed in Figure 1. The average UPC level for the reference group (defined as the estimated level for a 13-year-old non-African American girl with a non-glomerular diagnosis and at the previous visit presenting with an eGFR of 50 ml/min/1.73 m², UPC of 0.5 mg/mg Cr and SBP and DBP at the 50^th percentile) was 0.48 mg/mg Cr. In the absence of ACEi/ARB use, dhCCB use was significantly associated with higher proteinuria. Specifically, relative to those with no history of dhCCB use or ACEi/ARB use, those who had no history of ACEi/ARB use and used dhCCBs had 18.8% higher UPC levels (95% CI: +4.1%, +33.4%). In contrast, among ACEi/ARB users, there was no difference in UPC relative to those with no history of dhCCB or ACEi/ARB use, regardless of the pattern of dhCCB use. Those who discontinued ACEi/ARBs had higher levels of UPC overall. Specifically, those who discontinued ACEi/ARBs and did not use dhCCB had 50.3% (95% CI: +30.8%, +69.8%) higher UPC, and those who discontinued ACEi/ARBs and used dhCCB had 58.1% (95% CI: +29.7%, +86.5%) higher UPC. Three person-visits where participants discontinued both ACEi/ARBs and dhCCBs were excluded from the model due to low representation.

Fig. 1.

Adjusted percent differences in proteinuria with 95% confidence intervals by no dihydropyridine calcium channel blocker (dhCCB) use, current dhCCB use, discontinued dhCCB use, stratified by no ACEi/ARB use, current ACEi/ARB use, and discontinued ACEi/ARB use (n = 2503 person-visits). The reference group comprises person-visits reporting no dhCCB use and no ACEi/ARB use. Percent differences based on log-linear regression models with generalized estimating equations (GEE) to account for the dependencies of within-person repeated measurements. All models were adjusted for age, sex, race, underlying glomerular vs. non-glomerular diagnosis, previous GFR, previous systolic and diastolic blood pressure z-scores, and previous proteinuria level. Estimates in gray are based on less than 20 person-visits; levels among those who discontinued both ACEi/ARB and dhCCBs were not estimated (n.e.) due to less than 5 person-visits in this group

Figure 2 presents differences in SBP z-score by current and previous dhCCB use, stratified by ACEi/ARB use. Regardless of ACEi/ARB use, those who use dhCCBs had significantly higher SBP z-scores than those with no current of previous history of dhCCB or ACEi/ARB use: 0.38 (95% CI: +0.21, +0.54) SDs higher among those who use dhCCBs with no history of ACEi/ARB use, 0.32 (95% CI, 0.18 to 0.46) SDs higher among those who use dhCCBs and ACEi/ARBs, and 0.66 (95% CI, 0.38 to 0.94) SDs higher among those who use dhCCBs and discontinued ACEi/ARB use. There was no difference in SBP z-scores among those who discontinued dhCCBs, regardless of history of ACEi/ARB use. Discontinuation of ACEi/ARBs with no history of dhCCB use was associated with significantly higher SBP z-scores (+0.37 SDs; 95% CI, 0.16, 0.57) compared to the reference group.

Fig. 2.

Adjusted differences in systolic blood pressure z-score with 95% confidence intervals by no dihydropyridine calcium channel blocker (dhCCB) use, current dhCCB use, discontinued dhCCB use, stratified by no ACEi/ARB use, current ACEi/ARB use, and discontinued ACEi/ARB use (n = 2503 person-visits). The reference group comprises person-visits reporting no dhCCB use and no ACEi/ARB use. Differences based on log-linear regression models with generalized estimating equations (GEE) to account for the dependencies of within-person repeated measurements. All models were adjusted for age, sex, race, underlying glomerular vs. non-glomerular diagnosis, previous GFR, previous systolic and diastolic blood pressure z-scores, and previous proteinuria level. Estimates in gray are based on less than 20 person-visits; levels among those who discontinued both ACEi/ARB and dhCCBs were not estimated (n.e.) due to less than 5 person-visits in this group.

Figure 3 displays the differences in DBP z-scores from the same model structure as above. Similar to the results with SBP z-scores as the outcome, dhCCB use in the absence of ACEi/ARB use was associated with significantly higher DBP z-scores. Specifically, relative to those with no history of dhCCB or ACEi/ARB use, dhCCB use was significantly associated with higher levels of DBP z-scores among those with no previous or current ACEi/ARB use (0.26 SDs; 95% CI, 0.10, 0.43) and those who discontinued using ACEi/ARBs (0.51 SDs; 95% CI, 0.19, 0.83). This effect was essentially null among those who report concurrent ACEi/ARB use: ACEi/ARB users, in the absence of dhCCB use, had lower DBP z-scores (-0.15 SDs; 95% CI,-0.23, -0.07). Participants who discontinued ACEi/ARBs without a history of dhCCB use had 0.31 SD higher DBP z-scores (95% CI, 0.13, 0.49).

Fig. 3.

Adjusted differences in diastolic blood pressure z-score with 95% confidence intervals by no dihydropyridine calcium channel blocker (dhCCB) use, current dhCCB use, discontinued dhCCB use, stratified by no ACEi/ARB use, current ACEi/ARB use, and discontinued ACEi/ARB use (n = 2503 person-visits). The reference group comprises person-visits reporting no dhCCB use and no ACEi/ARB use. Differences based on log-linear regression models with generalized estimating equations (GEE) to account for the dependencies of within-person repeated measurements. All models were adjusted for age, sex, race, underlying glomerular vs. non-glomerular diagnosis, previous GFR, previous systolic and diastolic blood pressure z-scores, and previous proteinuria level. Estimates in gray are based on less than 20 person-visits; levels among those who discontinued both ACEi/ARB and dhCCBs were not estimated (n.e.) due to less than 5 person-visits in this group

Discussion

This study demonstrates an increase in proteinuria among patients on L-type dhCCBs in the absence of ACEi/ARB therapy. Not only was this increase not seen in those receiving ACEi/ARB therapy but also discontinuation of ACEi/ARB use led to slight worsening of proteinuria in patients on dhCCB therapy. These findings are significant because proteinuria and uncontrolled hypertension have been identified as critical risk factors for progression in pediatric CKD. Fathallah-Shaykh et al. demonstrated in 522 patients with non-glomerular CKD, that a more rapid decline in GFR of 0.3 ml/min/1.73 m² (95%CI 0.4, 0.1) per year was associated with a twofold higher baseline urine protein to creatinine ratio [21]. A higher baseline systolic blood pressure was also associated with a 0.4 ml/min/1.73 m² greater decline in GFR [21]. Analyses have also shown that RAAS use and strict BP control reduced the risk of progression to kidney replacement therapy from between 21 to 37% [4, 22]. However, in patients who discontinued RAAS blockade, these subjects were more likely to have a lower GFR, higher UPC, and higher blood pressures [21]. The findings in the current study, in which those who discontinued ACEi/ARB experienced a subsequent increase in UPC, support these conclusions.

Studies conducted in adults, such as the Irbesartan Diabetic Nephropathy Trial and the African American study of Kidney Disease and Hypertension, have demonstrated an increase in proteinuria associated with amlodipine use [11, 12]. A small randomized trial of 100 patients with CKD treated with manidipine (a L-T-dhCCB) versus nifedipine noted a 25% increase in proteinuria in patients treated with nifedipine [23]. However, a systematic review conducted by Bakris et al. only demonstrated a 1% increase in proteinuria with dihydropyridine calcium antagonists compared to a 26% reduction with use of non-dihydropyridine calcium antagonists [24]. Several recent trials comparing treatment with amlodipine (an L-type dhCCB) to L/T- or L/N-type dhCCBs did not note significant changes in proteinuria over time, but these studies were under-powered to address these issues specifically [25, 26]. Interestingly, the addition of L/T and L/N CCBs such as benidipine or cilnidipine was associated with improvement in proteinuria and preservation of kidney function [27].

These findings provide a potential mechanism for the negative effects of L-type dhCCBs on proteinuria due to the differential expression of voltage-dependent calcium channels within the kidney. It is postulated that L-type dhCCBs such as amlodipine exert their effects solely on the afferent arterioles leading to attenuation of autoregulation and transmission of systemic pressures [25–28]. The resultant glomerular hypertension induces enhanced protein filtration, endothelial damage, and fibrosis. In contrast, T- and N-type channels are expressed in both afferent and efferent arterioles, therefore, minimizing glomerular hypertension and preventing proteinuria as well as any reduction in GFR [29, 30].

We also found higher SBP and DBP among individuals using L-type dhCCB therapy, particularly in patients without current or prior use of ACEi/ARBs. However, we advise caution in interpreting these data as it is not possible to determine if this finding is due to higher baseline blood pressures or failure of dhCCB therapy to improve blood pressure control. Amlodipine, the most common dhCCB utilized in this population, has been found to be safe and effective in pediatric hypertension in both a short-term clinical trial and a long-term retrospective case series [31, 32]. Our findings suggest that the use of dhCCBs in this patient population requires further study. We note that the lack of BP improvement observed among those receiving dhCCBs could be due to being resistant to ACEi/ARBs, having hypertension that is more difficult to control, or having generally poor medication adherence. Additional investigation is warranted.

A major challenge in this analysis was dealing with potential confounding by indication, such that those who do not respond to ACEi/ARBs or are non-compliant with more severe disease may be more likely to be prescribed dhCCBs as an alternative. Fortunately, the longitudinal design of CKiD allowed us to control for disease severity based on the previous annual visit at which dhCCB use was assessed. While we recognize confounding by unmeasured indications may be a source of bias, our approach allowed us to adjust for key risk factors of severity, including proteinuria level at the prior year.

Our study was limited by the observational nature of our dataset. The control group was a mixture of children who were normotensive and hypertensive. The ideal comparison group would have been children who were hypertensive and could have received dhCCBs but did not; however, the study population had limited numbers of this target population, and data were not collected on indications for dhCCBs. We attempted to correct for variable imbalances through covariate adjustment, but it would have been ideal to have a dataset restricted to those most likely to receive dhCCBs. We also did not have information on prior dhCCB use among prevalent users and reasons for discontinuation. dhCCB use was relatively sparse and there was not enough data to characterize use beyond two visits (i.e., a visit pair). While we did explore differences between incident dhCCB use and persistent (or continued) dhCCB use, there were no significant differences between these two groups in terms of proteinuria differences and they were classified as the same in this analysis (“dhCCB users”), although we acknowledge this limitation in classification. Most of the participants receiving dhCCB in this study were prescribed amlodipine, which limits our ability to compare various dhCCBs. Data regarding dosage of dhCCB was also not available in this dataset. Lastly, these results may not be generalizable to children with higher GFRs who use dhCCBs.

This study is the first to report that use of dhCCBs in the absence of RAAS inhibition is associated with increased proteinuria in pediatric patients with CKD. Because of the importance of controlling proteinuria and hypertension to prevent progression of CKD, these results highlight the importance of ACEi and ARBs as first line agents in patients with pediatric CKD and may suggest that dhCCBs should be used with caution, or at least not without concurrent ACEI/ARB therapy. These findings warrant additional investigation to clarify the association between dhCCBs and proteinuria in CKD.

Data Availability

All data generated or analyzed during this study are included in this published article (and its supplementary information files).