Albuminuria, Proteinuria, and Renal Disease Progression in Children with CKD

Abstract

Background and objectives

The role of albuminuria as an indicator of progression has not been investigated in children with CKD in the absence of diabetes.

Design, setting, participants, & measurements

Children were enrolled from 49 centers of the CKD in Children study between January of 2005 and March of 2014. Cross-sectional multivariable linear regression (n=647) was used to examine the relationship between urine protein-to-creatinine (UP/C [milligrams per milligram]) and albumin-to-creatinine (ACR [milligrams per gram]) with eGFR (milliliters per minute per 1.73 m²). Parametric time-to-event analysis (n=751) was used to assess the association of UP/C, ACR, and urine nonalbumin-to-creatinine (Unon-alb/cr [milligrams per gram]) on the time to the composite endpoint of initiation of RRT or 50% decline in eGFR.

Results

The median follow-up time was 3.4 years and 202 individuals experienced the event. Participants with a UP/C≥0.2 mg/mg and ACR≥30 mg/g had a mean eGFR that was 16 ml/min per 1.73 m² lower than those with a UP/C<0.2 mg/mg and ACR<30 mg/g. Individuals with ACR<30 mg/g, but a UP/C≥0.2 mg/mg, had a mean eGFR that was 9.3 ml/min per 1.73 m² lower than those with a UP/C<0.2 mg/mg and ACR<30 mg/g. When categories of ACR and Unon-alb/cr were created on the basis of clinically meaningful cutoff values of UP/C with the same sample sizes for comparison, the relative times (RTs) to the composite end-point were almost identical when comparing the middle (RT=0.31 for UP/C [0.2–2.0 mg/mg], RT=0.38 for ACR [56–1333 mg/g], RT=0.31 for Unon-alb/cr [118–715 mg/g]) and the highest (RT=0.08 for UP/C [>2.0 mg/mg], RT=0.09 for ACR [>1333 mg/g], RT=0.07 for Unon-alb/cr [>715 mg/g]) levels to the lowest levels. A similar trend was seen when categories were created on the basis of clinically meaningful cutoff values of ACR (<30, 30–300, >300 mg/g).

Conclusions

In children with CKD without diabetes, the utility of an initial UP/C, ACR, and Unon-alb/cr for characterizing progression is similar.

Introduction

Proteinuria is a major prognostic indicator of renal progression in both children and adults with CKD (1–4). The Chronic Kidney Disease in Children (CKiD) study reported that total urine protein-to-creatinine ratios (UP/C) >2.0 mg/mg in children with glomerular CKD were associated with a 94% reduction in the time to either a 50% decline in eGFR or the initiation of RRT (1). Children with nonglomerular CKD and a UP/C>2.0 mg/mg had a 79% reduction in the time to this same endpoint. Indeed, compared with hypoalbuminemia, elevated BP, dyslipidemia, and anemia, nephrotic range proteinuria was the strongest risk factor for renal progression in children with CKD, regardless of glomerular or nonglomerular cause (1). In 2009, the randomized, prospective The Effect of Strict Blood Pressure Control and ACE Inhibition on the Progression of Chronic Renal Failure in Pediatric Patients (ESCAPE) trial in children with CKD demonstrated that higher levels of proteinuria were associated with a more rapid decline in GFR (3).

The 2012 Kidney Disease Improving Global Outcomes guidelines for staging and predicting the progression of CKD designate a urine albumin-to-creatinine ratio (ACR) of <30 mg/g as “normal-to-mildly increased,” 30–300 mg/g as “moderately increased,” and >300 mg/g as “severely increased.” (5) Traditionally, in studies of adults with CKD, albuminuria, rather than overall proteinuria, has been utilized in studies of progression (2, 6–10). Although the association of normal or mildly increased levels of albuminuria on worsening renal function has been explored in the adult CKD population, the association of smaller amounts of albuminuria (alternatively termed “microalbuminuria”) and kidney disease progression in pediatric patients has largely been confined to studies of patients with diabetes mellitus (8, 11). Because of this difference in urine protein quantification methods in adult versus pediatric studies, we wanted to determine whether albuminuria, specifically, has particular importance compared with general proteinuria as an indicator of renal progression in children without diabetes mellitus.

Evaluation for albuminuria is currently not part of the routine care of children with CKD without diabetes. Likewise, investigating the amount of nonalbumin proteins in the urine, whose increased presence indicates tubular dysfunction, is not routinely done in the evaluation of pediatric patients with CKD (12–14). This study sought to: (1) quantify the cross-sectional relationship between both UP/C and ACR with eGFR, (2) determine the cross-sectional relationships between UP/C, ACR, and urine nonalbumin-to-creatinine (Unon-alb/cr), and (3) compare the association of each of UP/C, ACR, and Unon-alb/cr with time to RRT or >50% decline in eGFR.

Materials and Methods

Study Population

The CKiD study is a multicenter, prospective cohort study of children with mild-to-moderate CKD conducted at 49 nephrology centers across North America. The study design and conduct were approved by an observational study-monitoring board appointed by the National Institute of Diabetes and Digestive and Kidney Diseases and by the internal review boards of each participating center. Each participating family provided informed consent. The demographic and clinical characteristics of the cohort as a whole have been published elsewhere (15). Study participants were seen at annual follow-up visits after their initial baseline visit which occurred between January of 2005 and March of 2014. Beginning in June of 2008, urine albumin was added to the urine panel of tests as part of each annual visit. Of the 891 participants enrolled, 757 had at least one visit with ACR measured; the visit with the first ACR measurement was defined as the index visit. Our time-to-event analyses were restricted to the 751 participants who had follow-up data after the index visit to determine time to first of >50% decline in eGFR or RRT. eGFR (milliliters per minute per 1.73 m²) was calculated at each visit using: 0.413×(height [centimeters]/serum creatinine [milligrams per deciliter]) (16).

Protein Measures

Participants provided a random urine collection on the morning of the study. Urine albumin was measured using an immunoturbidimetric assay whereby anti-albumin antibodies react with the urine sample to form antigen-antibody complexes. After agglutination, the antigen-antibody complexes were measured turbidimetrically by quantifying the loss of intensity of a light beam placed through the solution. The quantity of albumin (milligrams) was then divided by the quantity of creatinine (grams) in the urine in order to obtain the ratio, ACR. The total urine protein content of the urine was also determined by a turbidimetric method. The quantity of protein (milligrams) was divided by the quantity of creatinine (milligrams) in the urine to obtain the ratio, UP/C. Unon-alb/cr was defined as urine total protein (milligrams) minus urine albumin (milligrams) divided by creatinine (grams). Coefficients of variation on the Roche Hitachi Cobas modular analyzer for within-run and between-run assays were 1.5% and 3.5%, respectively, for urine albumin, and 1.5% and 1.5% for urine protein.

Statistical Analyses

Linear regression was used to examine the cross-sectional relationship between mean eGFR level and both UP/C and ACR in study participants in whom either proteinuria level was elevated in comparison to the reference group (UP/C<0.2 mg/mg and ACR<30 mg/g) using the following categories: UP/C<0.2 mg/mg and ACR≥30 mg/g; UP/C≥0.2 mg/mg and ACR<30 mg/g; UP/C≥0.2 mg/mg and ACR≥30 mg/g (5, 17–19), overall, and also among those not reporting an angiotensin converting enzyme (ACE) inhibitor or angiotensin receptor blocker (ARB) at the index visit. On the basis of previous studies showing significant associations with eGFR, the linear regression model was adjusted for: age, sex, race (white versus nonwhite), CKD cause (glomerular versus nonglomerular), hypertension status (systolic or diastolic BP ≥age-sex-height–specific 95th percentile, versus systolic and diastolic BP <age-sex-height–specific 95th percentile), age-sex–specific body mass index (≥95th percentile versus <95th percentile), and uric acid values (1, 20–22). Pearson correlation coefficients were used to quantify the strength of the cross-sectional linear relationship between log(UP/C) and log(ACR), log(UP/C) and log(Unon-alb/cr), and log(ACR) and log(Unon-alb/cr) at the index visit.

Parametric failure time models assuming a generalized gamma (GG) distribution (23) of event times were used to assess the independent association of index visit values of UP/C, ACR, and Unon-alb/cr on the time to the composite end-point of either >50% decline in eGFR (compared with eGFR at the index visit) or initiation of RRT. Two sets of analyses were performed: one on the basis of clinically meaningful UP/C cutoffs and a second set on the basis of clinically meaningful ACR cutoffs. Because proteinuria is considered clinically relevant when the UP/C is ≥0.2 mg/mg and nephrotic range proteinuria is typically defined as a UP/C>2.0 mg/mg, these values were used as cutoffs in determining categories for UP/C (<0.2, 0.2–2.0, and >2.0 mg/mg) (19, 24). We then examined the distribution of ACR and Unon-alb/cr and chose cutoffs for these variables (<56, 56–1333, and >1333 mg/g for ACR; <118, 118–715, and >715 mg/g for Unon-alb/cr) that would yield the same sample sizes (287, 363, and 101) for the three different categories across the three sets of analyses. Similarly, because ACR is considered moderately increased when ≥30 mg/g and severely increased when >300 mg/g, these values were used as cutoffs for ACR categories (<30, 30–300, and >300 mg/g). Cutoffs for UP/C (<0.139, 0.139–0.630, and >0.630 mg/mg) and for Unon-alb/cr (<86, 86–273, and >273 mg/g) were chosen to yield the same sample sizes (220, 271, and 260) for the three different categories across the three sets of analyses.

Relative times (RTs) were used to quantify the strength of the relationship between each of these variables and the composite event. Specifically, RTs represent the time it takes for P% (P can be any percentage between 0% and 100%) of the “exposed group” to develop the composite event divided by the time it takes for the same P% of the “unexposed group” to develop the composite event. We chose to use RTs as opposed to relative hazards because time is the unit of measurement of the dependent variable, and to quantify the association of UP/C, ACR, and Unon-alb/cr on CKD progression in terms of how higher levels of each exposure shorten the time to the event (i.e., RT<1). We utilized regression models under the assumption of proportional times (i.e., the RTs did not depend on “P”). The appropriateness of the GG distribution was assessed by comparing each estimated survival curve on the basis of the GG parameter estimates with the corresponding nonparametric Kaplan–Meier survival curve. All analyses used a 0.05 level of significance.

Results

Characteristics at the index visit of the 751 CKiD participants included in our time-to-event analyses are shown in Table 1. The median age was 12.4 years, 15% of participants had stage 1 hypertension or greater, and the median eGFR was 54.6 ml/min per 1.73 m². A glomerular cause for CKD occurred in 30% of the participants. At the index visit, 56% of participants were on an ACE inhibitor or ARB.

Table 1.

Characteristics at the index visit of the 751 CKiD participants

Characteristics	Median (Interquartile Range) or % (n)
Age, yr	12.4 (8.6–15.7)
Men	61 (457)
White race	65 (491)
Systolic or diastolic BP ≥95th percentilea	15 (108)
Current ACE or ARB use	56 (417)
eGFRb, ml/min per 1.73 m2	54.6 (39.5–71.8)
Glomerular CKD cause	30 (228)
Age-sex–specific body mass index ≥95th percentilec	17 (129)
Urine protein-to-creatinine, mg/mg	0.32 (0.12–1.04)
Urine albumin-to-creatinine, mg/g	112.0 (20.9–615.0)
Urine nonalbumin-to-creatinine, mg/g	167.6 (75.0–400.7)
Urine albumin-to-protein, mg/mg	0.44 (0.15–0.65)

n, number of participants.

^aBP percentiles were determined using age/sex/height–specific values; missing for 33 participants.

^beGFR=0.413×(height [centimeters]/serum creatinine [milligrams per deciliter]).

^cBody mass index percentiles were determined using age/sex–specific values; missing for 12 participants.

Cross-Sectional Analyses

Table 2 shows the results of a multivariable linear regression analysis to quantify the relationship between mean eGFR level and both UP/C and ACR at the index visit. The mean eGFR in the individuals with albuminuria (ACR≥30 mg/g) in the absence of overt proteinuria (UP/C<0.2 mg/mg) did not differ significantly from those of the reference group (UP/C<0.2 mg/mg and ACR<30 mg/g). In contrast, those with both albuminuria and overt proteinuria (UP/C≥0.2 mg/mg and ACR≥30 mg/g) had a mean eGFR that was 16.0 ml/min per 1.73 m² lower (95% confidence interval [95% CI], −19.7 to −12.4) than that of the reference group. Importantly, participants with overt proteinuria (UP/C≥0.2 mg/mg), but ACR<30 mg/g, had a mean eGFR that was 9.3 ml/min per 1.73 m² lower (95% CI, −17.8 to −0.7) than those with UP/C<0.2 mg/mg and ACR<30 mg/g.

Table 2.

Multivariable^a linear regression analysis of index visit values of eGFR on urine protein-to-creatinine (milligrams per milligram) and urine albumin-to-creatinine (milligrams per gram), n=647

Exposureb	No.	Mean eGFR (millilitre per minute per 1.73 m2) Relative to Reference	95% Confidence Interval
Urine protein-to-creatinine <0.2 and urine albumin-to-creatinine <30	170	0 (reference)
Urine protein-to-creatinine <0.2 and urine albumin-to-creatinine ≥30	86	−3.1	−8.1 to 1.9
Urine protein-to-creatinine ≥0.2 and urine albumin-to-creatinine <30	21	−9.3	−17.8 to −0.7
Urine protein-to-creatinine ≥0.2 and urine albumin-to-creatinine ≥30	370	−16.0	−19.7 to −12.4

^aAdjusted for age, sex, race, CKD cause, hypertension status, body mass index, and uric acid.

^bFor urine protein-to-creatinine the units are mg/mg; for urine albumin-to-creatinine the units are mg/g.

We repeated the cross-sectional multivariable linear regression including only the 289 patients not on an ACE and/or ARB at the index visit (Supplemental Table 1). Similar results were obtained to the analysis where patients on an ACE and/or ARB were included; individuals with albuminuria (ACR≥30 mg/g) in the absence of overt proteinuria (UP/C<0.2 mg/mg) had a mean eGFR that was not significantly different from the reference group. In comparing the analyses with and without the restriction of ACE inhibitor/ARB use, the magnitude of the difference in means was greater when comparing those with UP/C≥0.2 mg/mg and ACR≥30 mg/g to those with UP/C<0.2 mg/mg and ACR<30 mg/g when the analysis was restricted to those not on an ACE or ARB.

Figure 1 shows the strong collinearity of ACR and UP/C (r=0.92), Unon-alb/cr and UP/C (r=0.92), and Unon-alb/cr and ACR (r=0.74) at the index visit.

Figure 1.

Collinearity between the three methods to quantify proteinuria at the index study visit, n=751. The dashed line shown in each of the panels in Figure 1 is the line on which all of the data would fall if there were perfect agreement between the two variables.

Time-to-Event Analyses

The median follow-up time from the index visit to the composite event was 3.4 years (interquartile range, 2.2–5.0) and 27% of the study population (n=202) were observed to develop the composite event (>50% decline in eGFR or initiation of RRT). Seventy-seven children had a >50% decline in eGFR before undergoing dialysis or transplant and 37 had a >50% decline in eGFR without subsequent dialysis or transplant. Eighty-eight children went onto RRT without first exhibiting a >50% decline in eGFR. As shown in Table 3, the times to the composite event were 92% shorter for those with UP/C>2.0 mg/mg when compared with participants with a UP/C<0.2 mg/mg (RT=0.08; 95% CI, 0.04 to 0.13). For albuminuria, the times to the composite event were 84% shorter for participants with ACR>300 mg/g when compared with those with ACR<30 mg/g (RT=0.16; 95% CI, 0.10 to 0.26). Similarly, the times to the composite event were 86% shorter for participants with a Unon-alb/cr≥300 mg/g when compared with Unon-alb/cr<100 mg/g (RT=0.14; 95% CI, 0.08 to 0.24).

Table 3.

Relationship between indicated exposure at index visit and subsequent decline of >50% in GFR or initiation of RRT, n=751

Exposure	No.	No. (%) with Event	Regression Modela	Relative Time	95% Confidence Interval
Urine protein-to-creatinine, mg/mg
<0.2	287	25 (9)	GG (3.553,1.086,0.690)	1 (reference)
0.2–2.0	363	112 (31)	GG (2.393,1.086,0.690)	0.31	0.21 to 0.47
>2.0	101	65 (64)	GG (0.974,1.086,0.690)	0.08	0.04 to 0.13
Urine albumin-to-creatinineb, mg/g
<56	287	31 (11)	GG (3.391,1.024,0.784)	1 (reference)
56–1333	363	104 (29)	GG (2.428,1.024,0.784)	0.38	0.26 to 0.55
>1333	101	67 (66)	GG (1.031,1.024,0.784)	0.09	0.06 to 0.15
Urine nonalbumin-to-creatinineb, mg/g
<118	287	25 (9)	GG (3.593,1.233,0.507)	1 (reference)
118–715	363	112 (31)	GG (2.411,1.233,0.507)	0.31	0.20 to 0.46
>715	101	65 (64)	GG (0.936,1.233,0.507)	0.07	0.04 to 0.12

GG, generalized gamma.

^aIncludes the location, scale, and shape parameters of the regression model.

^bCutoffs for urine albumin-to-creatinine and urine nonalbumin-to-creatinine were chosen to yield same sample sizes as determined by urine protein-to-creatinine cutoffs of 0.2 and 2.0.

In order to compare the three methods to quantify proteinuria, UP/C, ACR, and Unon-alb/cr were stratified into categories on the basis of levels of UP/C (<0.2 mg/mg [n=287], 0.2–2.0 mg/mg [n=363], and >2.0 mg/mg [n=101]) as shown (Figure 2, Table 3). Categories of ACR and Unon-alb/cr were chosen on the basis of where 0.2 and 2.0 fell in the distribution of UP/C so that the lowest levels of UP/C, ACR, and Unon-alb/cr would have the same sample sizes but not necessarily the same participants. Similarly, the middle categories (as well as the highest categories) of UP/C, ACR, and Unon-alb/cr would have the same number of participants. The lowest category of each measure of proteinuria was designated as a reference group (i.e., UP/C<0.2 mg/mg, ACR<56 mg/g, and Unon-alb/cr<118 mg/g). The RTs to a >50% decline in eGFR or the initiation of RRT were almost identical for each of the three measures of proteinuria when comparing the middle category to the reference group (0.31 for UP/C, 0.38 for ACR, 0.31 for Unon-alb/cr) and highest category to the reference group (0.08 for UP/C, 0.09 for ACR, 0.07 for Unon-alb/cr).

Figure 2.

Similar prognostic ability for each of the three methods to quantify proteinuria for characterizing a >50% decline in GFR or need for RRT based on clinically meaningful cutoffs of UP/C (Kaplan–Meier, dashed lines and Generalized Gamma, solid lines). Left panel: UP/C categorized as <0.2, 0.2–2.0, and >2.0; middle panel: ACR categorized as <56, 56 to <1333, and >1333; right panel: Unon-alb/cr categorized as <118, 118 to <715, and >715. n=751. ACR, urine albumin-to-creatinine ratio; Unon-alb/cr, urine nonalbumin-to-creatinine ratio; UP/C, urine protein-to-creatinine ratio.

In order to further explore the subgroup of study participants with UP/C<0.2 mg/mg (n=287), the association between ACR (treated as a continuous variable) and the composite event was examined. Each log higher in ACR was not associated with any greater risk of reaching the composite endpoint (P=0.96).

When UP/C, ACR, and Unon-alb/cr were compared on the basis of the levels of ACR (<30 mg/g [n=220], 30–300 mg/g [n=271], and >300 mg/g [n=260]), the RTs to the event were very similar when comparing the middle and highest categories of each of the three measures of proteinuria to the reference group (Figure 3, Table 4).

Figure 3.

Similar prognostic ability for each of the three methods to quantify proteinuria for characterizing a >50% decline in GFR or need for RRT based on clinically meaningful cutoffs of ACR (Kaplan–Meier, dashed lines and Generalized Gamma, solid lines). Left panel: UP/C categorized as <0.139, 0.139–0.630, and >0.630; middle panel: ACR categorized as <30, 30–300, and >300; right panel: Unon-alb/cr categorized as <86, 86–273, and >273. n=751. ACR, urine albumin-to-creatinine ratio; Unon-alb/cr, urine nonalbumin-to-creatinine ratio; UP/C, urine protein-to-creatinine ratio.

Table 4.

Relationship between indicated exposure at index visit and subsequent decline of >50% in GFR or initiation of RRT, n=751

Exposure	No.	No. (%) with Event	Regression Modela	Relative Time	95% Confidence Interval
Urine protein-to-creatinineb, mg/mg
<0.139	220	17 (8)	GG (3.721,0.935,0.991)	1 (reference)
0.139–0.630	271	46 (17)	GG (3.043,0.935,0.991)	0.51	0.30 to 0.86
>0.630	260	139 (53)	GG (1.662,0.935,0.991)	0.13	0.07 to 0.22
Urine albumin-to-creatinine, mg/g
<30	220	23 (10)	GG (3.469,0.932,0.977)	1 (reference)
30–300	271	46 (17)	GG (3.001,0.932,0.977)	0.63	0.40 to 1.00
>300	260	133 (51)	GG (1.663,0.932,0.977)	0.16	0.10 to 0.26
Urine nonalbumin-to-creatinineb, mg/g
<86	220	18 (8)	GG (3.913,0.336,3.328)	1 (reference)
86–273	271	49 (18)	GG (3.166,0.336,3.328)	0.47	0.27 to 0.84
>273	260	135 (52)	GG (2.065,0.336,3.328)	0.16	0.09 to 0.28

GG, generalized gamma.

^aIncludes the location, scale, and shape parameters of the regression model.

^bCutoffs for urine protein-to-creatinine and urine nonalbumin-to-creatinine were chosen to yield same sample sizes as determined by urine albumin-to-creatinine cutoffs of 30 and 300.

Discussion

In this large study of North American children with CKD without diabetes, it was found that UP/C, ACR, and Unon-alb/cr had a similar ability to characterize progression to a composite event of a >50% decline in eGFR or initiation of RRT. The times to the event were significantly shorter for increasing levels of UP/C, ACR, and Unon-alb/cr. Consistent with previous studies, higher UP/C values were significantly associated with lower eGFR values (1–4). Not unexpectedly, when the UP/C was ≥0.2 mg/mg, there was a significant association with lower mean eGFR values for individuals with ACR≥30 mg/g (i.e., those with UP/C≥0.2 mg/mg and ACR≥30 mg/g had significantly lower eGFR compared with those with UP/C<0.2 mg/mg and ACR<30 mg/g). Importantly, however, a significant association with lower eGFR was also seen in participants with a UP/C≥0.2 mg/mg and ACR<30 mg/g.

To our knowledge, this study is the first to examine the association of albuminuria and renal progression in a pediatric CKD population in the absence of diabetes. The results of this analysis indicate that ACR and UP/C are very well correlated (r=0.92) and the discrepancy in their correlation does not yield differential inferences with respect to characterizing the time to >50% decline in eGFR or initiation of RRT in the population studied. The congruence of the results for the analyses whereby methods of protein quantification were compared on the basis of UP/C as well as ACR categories indicates an equivalent prognostic ability for each of these markers in terms of characterizing CKD progression.

It can be argued that screening for albuminuria, which is more sensitive than the assay for proteinuria, may detect patients with CKD at an earlier point in their chronic disease. Prior investigations have supported the belief that microalbuminuria in patients with diabetes provides an ominous connotation regarding the progression of the diabetes (25–27). However, more recent studies have shown that microalbuminuria may be a transient finding in children and adults with diabetes (28, 29) and is not consistently a strong predictor of renal disease progression in this patient population (30–32). The cost of measuring urine albumin is more than the cost of measuring total protein. Thus, there may be a cost-saving benefit to avoiding additional urine protein studies when UP/C is already known. However, future cost effectiveness studies are required to make any determinations of the cost-saving advantage of one method of urine protein quantification over another in the pediatric CKD population.

Although our study demonstrates that albuminuria is not superior to proteinuria in monitoring renal progression in children with CKD, there may be utility in quantifying both a urine protein and albumin in certain patient groups. Prior studies in both children and adults with nephrotic syndrome have shown a significant association of the selectivity of the proteinuria and prognosis and response to therapy (33–36). In order to further explore the role of urine albumin in this study cohort, UP/C, ACR, and Unon-alb/cr were dichotomized on the basis of a urine albumin-to-protein ratio (APR [milligrams per milligram] of <0.4 and ≥0.4). The 0.4 cutoff value was on the basis of prior study results showing the utility of an APR for identifying tubular interstitial disorders in adults (13). The RTs to a >50% decline or initiation of RRT comparing participants with the highest APR (>0.4 mg/mg), highest UP/C (>0.28 mg/mg), highest ACR(>90 mg/g), and highest Unon-alb/cr (>148 mg/g) compared with those with the lowest were 0.44, 0.20, 0.23, and 0.20, respectively, suggesting that obtaining an APR may provide somewhat different information regarding progression when compared with UP/C, ACR, and Unon-alb/Cr. Further studies are needed to determine if APR is a useful measurement in children with CKD.

The study does have limitations. A single random urine specimen was obtained on each participant, not accounting for potential intraindividual variability in urine protein content. Additionally, although urine collections were obtained in the morning of the study visit, specimens were not consistently first morning voids, providing the potential for orthostatic proteinuria. Nonalbumin protein was calculated indirectly using urine total protein and albumin measurements. We did not directly measure specific larger–molecular weight tubular proteins, which may have added utility in predicting CKD progression in children. In adults with primary GN, the fractional excretion of both IgG and albumin have been found to be significant predictors for the progression to stage 5 CKD or death (33). Notably, albuminuria has been shown in multiple studies to be an independent indicator of increased cardiovascular disease risk (37–40), an outcome not explored in this study. The exact timing of ACE inhibitor and/or ARB medication initiation could not be determined from the study data. Given that these medications are common among the CKD population and can result in a stabilization or decline in the degree of proteinuria and albuminuria (18, 41), patients not on an ACE inhibitor or ARB at the index study visit were included in a cross-sectional multivariable linear regression with similar results to the entire patient cohort.

In conclusion, in children with CKD without diabetes, initial UP/C, ACR, and Unon-alb/cr determinations are well correlated and provide a similar characterization of CKD progression.

Disclosures

G.J.S. is a consultant for Tricida. The remaining coauthors have nothing to disclose.