Research

JAMA Pediatrics | Original Investigation

Use of the Kidney Failure Risk Equation to Determine

the Risk of Progression to End-stage Renal Disease
in Children With Chronic Kidney Disease
Erica Winnicki, MD; Charles E. McCulloch, PhD; Mark M. Mitsnefes, MD; Susan L. Furth, MD, PhD;
Bradley A. Warady, MD; Elaine Ku, MD

Editorial page 122

IMPORTANCE The kidney failure risk equation (KFRE) has been shown to accurately estimate
progression to kidney failure in adults with chronic kidney disease (CKD). Use of the KFRE in
children with CKD, if accurate, would help to optimize planning for end-stage renal disease
(ESRD) care.

OBJECTIVE To determine whether the KFRE adequately discriminates the risk of ESRD in
children with CKD.

DESIGN, SETTING, AND PARTICIPANTS This retrospective cohort study included 603 children
with an estimated glomerular filtration rate of less than 60 mL/min/1.73 m2 in the Chronic
Kidney Disease in Children study, a national multicenter observational study. Data were
collected from January 1, 2005, through July 31, 2013, and analyzed from September 30,
2016, through September 8, 2017.

EXPOSURES The primary predictive factors were the 4-variable (age, sex, bedside Schwartz
estimated glomerular filtration rate, and ratio of albumin to creatinine levels) and 8-variable
(4 variables plus serum calcium, phosphate, bicarbonate, and albumin levels) KFREs, which
provide 1-, 2-, and 5-year estimates of the risk of progression to ESRD.

MAIN OUTCOMES AND MEASURES Time to ESRD. The Cox proportional hazards model was
used to examine the association between the KFRE score and time to ESRD. C statistics were
used to discriminate ESRD risk by the KFRE, with a value of greater than 0.80 indicating
strong discrimination.

RESULTS Of the 603 children included in the study, 378 were boys (62.7%) and 225 were girls
(37.3%); median age at study entry was 12 years (interquartile range, 8-15 years). Median Author Affiliations: Division of
Nephrology, Department of
estimated glomerular filtration rate was 44 mL/min/1.73 m2. Four hundred fifty-seven Pediatrics, University of California,
participants (75.8%) had a nonglomerular cause of CKD. Median observation time was 3.8 San Francisco (Winnicki, Ku);
years (interquartile range, 1.7-6.2 years); 144 (23.9%) developed ESRD within 5 years of Department of Epidemiology and
Biostatistics, University of California,
enrollment. The 4-variable KFRE scores discriminated risk of ESRD, with C statistics of 0.90,
San Francisco (McCulloch); Division
0.86, and 0.81 for the 1-, 2-, and 5-year risk scores, respectively. Results were similar using the of Nephrology and Hypertension,
8-variable equation. Department of Pediatrics, Cincinnati
Children’s Hospital, Cincinnati, Ohio
(Mitsnefes); Division of Nephrology,
CONCLUSIONS AND RELEVANCE The KFRE is a simple tool that provides excellent Department of Pediatrics, Children’s
discrimination of the risk of ESRD. Results suggest that the KFRE could be incorporated into Hospital of Philadelphia, Philadelphia,
the clinical care of children with CKD to aid in anticipatory guidance, timing of referral for Pennsylvania (Furth); Division of
Nephrology, Department of
transplant evaluation, and planning for dialysis access.
Pediatrics, Children’s Mercy Kansas
City, Kansas City, Missouri (Warady);
Division of Nephrology, Department
of Medicine, University of California,
San Francisco (Ku).
Corresponding Author: Erica
Winnicki, MD, Division of Nephrology,
Department of Pediatrics,
University of California, San
Francisco, 550 16th St, 5th Floor,
JAMA Pediatr. 2018;172(2):174-180. doi:10.1001/jamapediatrics.2017.4083 San Francisco, CA 94143 (erica
Published online December 18, 2017. .winnicki@ucsf.edu).

174 (Reprinted) jamapediatrics.com

© 2017 American Medical Association. All rights reserved.

Downloaded From: by a Sam Ratulangi University Hospital User on 04/24/2018

 Kidney Failure Risk Equation to Determine Risk of End-stage Renal Disease
T
he kidney failure risk equations (KFREs) were origi-
nally developed in a large Canadian cohort to aid in clini-
cal decision making for adults with chronic kidney dis-
ease (CKD).1 These equations have been shown to be highly
accurate in the identification of adults at high risk for progres-
sion to end-stage renal disease (ESRD) and can be used to help
optimize the timing of dialysis access and/or kidney trans-
plant. Since their development, these risk equations have been
validated in diverse adult populations worldwide.2-4
Although risk factors for disease progression in children with
CKD have been well-established,5 no established risk equations
currently help to determine when a child will likely require re-
nal replacement therapy. Improved estimation of when a child
with CKD will experience progression to ESRD has several po-
tentially valuable applications. First, knowledge of the risk of
ESRD within a specified period would improve the ability of
clinicians to give anticipatory guidance to patients and their
families. Second, routine use of the KFRE may improve rates of
preemptive kidney transplant (transplant before the need for
dialysis occurs). Preemptive transplant currently occurs in only
22% of children with ESRD despite increasing evidence in chil-
dren with CKD that preemptive transplant improves allograft and
patient survival6-9 and that prompt transplant confers additional
benefits, including improved health-related quality of life, cog-
nition, and growth in children, compared with maintenance
dialysis.10-13 Third, better estimation of the timing of ESRD
onset would allow for improved completion of immunizations
in children with CKD before transplant, which is currently sub-
optimal despite the significant morbidity and mortality associ-
ated with vaccine-preventable infections after transplant.14,15
In addition, in rural areas with no ready access to pediatric
nephrologists,16 improved estimation of when a child will prog-
ress to ESRD might aid general practitioners in timing referrals
to nephrology. The objective of this study was to apply the KFREs
to a pediatric CKD cohort to determine whether these equations
accurately discriminate the risk of ESRD in children.
Methods
Study Population
We studied participants enrolled in the Chronic Kidney Disease
in Children (CKiD) study.17 Details of the CKiD study design have
been published previously.18,19 In brief, the CKiD study enrolled
children and adolescents aged 1 to 16 years (hereinafter referred
to as children) with an estimated glomerular filtration rate
(eGFR) from 30 to 90 mL/min/1.73 m2 and followed them up pro-
spectively from January 1, 2005, through July 31, 2013. For the
purpose of this study, we included all children in the CKiD study
with an eGFR of less than 60 mL/min/1.73 m2 using the bedside
Schwartz equation20 at baseline enrollment. We chose to include
only children with an eGFR of less than 60 mL/min/1.73 m2,
because the KFRE was developed and validated in adults with
an eGFR below this level. For participants who did not have an
eGFR of less than 60 mL/min/1.73 m2 at the time of baseline
enrollment into the CKiD study, the first visit when the eGFR
fell below 60 mL/min/1.73 m2 was used as their baseline visit.
Participants with missing covariates needed for determination
jamapediatrics.com
© 2017 American Medical Association. All rights reserved.
Downloaded From: by a Sam Ratulangi University Hospital User on 04/24/2018
Original Investigation Research
Key Points
Question Does the kidney failure risk equation, which was
developed for adults with chronic kidney disease, accurately
determine the risk of progression to end-stage renal disease in
children with chronic kidney disease?
Findings In this cohort study of 603 children with chronic kidney
disease in the Chronic Kidney Disease in Children study, the kidney
failure risk equation score was associated with progression to
end-stage renal disease at 1, 2, and 5 years.
Meaning Use of the kidney failure risk equation in children with
chronic kidney disease may improve the ability to determine the
short- and intermediate-term risk for progression to end-stage
renal disease; this knowledge may improve the timing of
appropriate anticipatory guidance and preparation for kidney
transplant or dialysis.
of the KFRE were excluded. Participants were also excluded if
they had a diagnosis of hyperoxaluria, because these partici-
pants may start renal replacement at higher levels of renal func-
tion than the norm. The institutional review board of the Uni-
versity of California considered this study exempt from research
approval for human subjects. Informed consent was obtained
locally from study participants at all CKiD sites.
Variables
The primary predictive factor was the calculated risk of kidney
failure (percentage of risk of developing ESRD within a 1-, 2-, or
5-year period). The 4-variable (age, sex, eGFR, and ratio of albu-
min to creatinine levels [ACR]) and the 8-variable (4 variables plus
serum calcium, phosphate, bicarbonate, and albumin levels)
KFREs were used as primary predictive factors in separate
models.1,3,4 The variables needed to determine the risk, as de-
scribed above, were obtained from the baseline visit or the visit
when eGFR first decreased to below 60 mL/min/1.73 m2. Because
albuminuria was not uniformly measured in the CKiD study, the
ratio of protein to creatinine levels (PCR) was transformed to an
ACR using a published equation.3 We ascertained the adequacy
ofthisconversionbydeterminingthecorrelationbetweentheACR
and the PCR transformed to ACR (Spearman r = 0.9; P < .001).
The study outcome of interest was time to ESRD, defined
as receipt of long-term dialysis or kidney transplant (which-
ever came first). Additional covariates of interest included
sex, race, ethnicity, cause of CKD, systolic and diastolic blood
pressure measurements, and hemoglobin level.
Statistical Analysis
Data were analyzed from September 30, 2016, through Sep-
tember 8, 2017. For our analysis, Cox proportional hazards mod-
els were used to examine the association between estimated
risk of kidney failure and time to ESRD. Postestimation C sta-
tistics were used to determine the ability of the KFRE to dis-
criminate the risk of ESRD. Confidence intervals for the C sta-
tistics and their difference were determined using a bootstrap
approach with 500 repetitions. We also examined whether dis-
crimination of the KFRE was different among the following pre-
specified subgroups: age (≥12 vs <12 years), sex, race (white vs
nonwhite), ethnicity (Hispanic vs non-Hispanic), and cause of
(Reprinted) JAMA Pediatrics February 2018 Volume 172, Number 2 175
 Research Original Investigation Kidney Failure Risk Equation to Determine Risk of End-stage Renal Disease

Table 1. Baseline Characteristics of the Cohort Table 2. C Statistics for the 4- and 8-Variable KFRE Applied to the CKiD Cohort
Data KFRE Risk C Statistic (95% CI)
Characteristic (N = 603)
4-Variable
Demographic
1-y 0.90 (0.86-0.93)
Age, median (IQR), y 12 (8-15)
2-y 0.86 (0.81-0.90)
Male, No. (%) 378 (62.7)
5-y 0.81 (0.77-0.83)
Race, No. (%)
8-Variable
White 401 (66.5)
1-y 0.91 (0.87-0.94)
Black 109 (18.1)
2-y 0.87 (0.82-0.91)
Asian, Pacific Islander, or Native American 23 (3.8)
5-y 0.82 (0.78-0.85)
Other or mixed race 70 (11.6)
Hispanic ethnicity, No. (%) 97 (16.3) Abbreviations: CKiD, Chronic Kidney Disease in Children; KFRE, kidney failure
risk equation.
Cause of CKD, No. (%)
Nonglomerular 457 (75.8)
Glomerular 146 (24.2) Meier survival curves to each risk tertile to determine whether
Physical examination the estimated survival matched the actual survival. In addi-
Weight, median (IQR), kg 39.1 (24.5-55.9)
tion, we compared characteristics of participants with a KFRE
Height, median (IQR), cm 143.1 (122.2-160.4)
score of greater than 13% (representing the highest tertile) at
Systolic blood pressure
2 years with those with a KFRE risk score of 13% or less at 2
Mean (SD), mm Hg 108 (13)
years using the unpaired 2-tailed t test, χ2 test, or Wilcoxon rank
Percentile based on age, sex, and height, 65 (37-87)
median (IQR) sum test as appropriate.
Diastolic blood pressure
Mean (SD), mm Hg 66 (11) Sensitivity Analyses
Percentile based on age, sex, and height, 68 (45-88) To determine whether the ACR derived from mathematical
median (IQR)
conversion would yield results similar to actual ACR measure-
Laboratory
eGFR
ments, we compared risk discrimination by the 2-year,
Median (IQR), mL/min/1.73 m2 44 (33-53) 8-variable KFRE among the subset of the population that
30-59 mL/min/1.73 m2, No. (%) 511 (84.7) had measured and calculated ACR data available.
15-29 mL/min/1.73 m2, No. (%) 92 (15.3) All data were derived from the National Institute of Diabe-
Hemoglobin level, mean (SD), g/dL 12.7 (3.5) tes and Digestive and Kidney Diseases Central Repository in de-
Serum calcium level, mean (SD), mg/dL 9.5 (0.6) identified form, and data were administratively censored as of
Serum phosphorus level, mean (SD), mg/dL 4.5 (0.8) July 2013. Dates of birth in this data set were rounded to the near-
Serum albumin level, mean (SD), g/dL 4.3 (0.5) est year to protect patient identity. All data analyses were con-
Serum bicarbonate level, mean (SD), mEq/L 20.9 (3.4) ducted using STATA software (version 13; StataCorp).
Ratio of urine protein to creatinine levels, 506 (167-1333)
median (IQR), mg/g
Ratio of urine albumin to creatinine level
(calculated)
Median (IQR), mg/g 220 (71-588)
Results
<30 mg/g, No. (%) 46 (7.6) A total of 603 participants (378 boys [62.7%] and 225 girls [37.3%];
30-299 mg/g, No. (%) 299 (49.6) median age at study entry, 12 years; interquartile range [IQR], 8-15
≥300 mg/g, No. (%) 258 (42.8) years) met our inclusion criteria and were included for study. We
Outcome identified 636 children in the cohort with an eGFR of less than
Observation time from baseline visit, 3.8 (1.7-6.2) 60 mL/min/1.73 m2, of whom 31 were missing data necessary for
median (IQR), y
Progression to ESRD within 5 y 144 (23.9) the KFRE calculations, and 2 additional participants were ex-
from baseline visit, No. (%) cluded for a diagnosis of hyperoxaluria. Of the 603 children
Abbreviations: CKD, chronic kidney disease; eGFR, estimated glomerular remaining, 528 had an eGFR of less than 60 mL/min/1.73 m2 at
filtration rate; ESRD, end-stage renal disease; IQR, interquartile range. baseline enrollment into the CKiD, and an additional 75 children
SI conversion factors: To convert albumin to grams per liter, multiply by 10; were included at the time when their eGFR decreased to less than
bicarbonate to millimoles per liter, multiply by 1.0; calcium to millimoles per
60 mL/min/1.73 m2 at a CKiD follow-up visit.
liter, multiply by 0.25; and phosphorus to millimoles per liter, multiply by 0.323.
Demographic, baseline physical examination, and
laboratory data are shown in Table 1. The median eGFR was
CKD (glomerular vs nonglomerular). Participants who died 44 mL/min/1.73 m2 (interquartile range [IQR], 33-53 mL/min/
(n = 2) were censored from our analysis at the time of death; 1.73 m2). Four hundred one participants (66.5%) were white,
therefore, deaths were not treated as a competing risk given 506 (83.9%) were non-Hispanic, and 457 (75.8%) had a non-
their rare occurrence within this pediatric cohort. glomerular cause of CKD. During the follow-up period, 27
In secondary analyses to determine whether the model was (4.5%) progressed to ESRD by 1 year of follow-up; 62 (10.3%),
well-calibrated, we divided the cohort into risk tertiles based by 2 years; and 144 (23.9%), by 5 years. The median follow-up
on the 2-year, 8-variable KFRE scores, and we fit Kaplan- time for the cohort was 3.8 years (IQR, 1.7-6.2 years).

176 JAMA Pediatrics February 2018 Volume 172, Number 2 (Reprinted) jamapediatrics.com

© 2017 American Medical Association. All rights reserved.

Downloaded From: by a Sam Ratulangi University Hospital User on 04/24/2018

 Kidney Failure Risk Equation to Determine Risk of End-stage Renal Disease Original Investigation Research

Figure 1. C Statistics for the 2-Year 8-Variable Kidney Failure Risk Equation by Patient Characteristic

No. of
Characteristic Participants C Statistic (95% CI) Difference (95% CI)
Age, y
≥12 309 0.85 (0.79 to 0.90)
0.09 (0.03 to 0.15)
<12 294 0.94 (0.91 to 0.97)
Race
White 401 0.87 (0.78 to 0.92)
0.01 (–0.09 to 0.07)
Nonwhite 202 0.88 (0.83 to 0.93)
Ethnicity
Hispanic 97 0.96 (0.91 to 0.98)
0.10 (0.04 to 0.16)
Non-Hispanic 499 0.86 (0.81 to 0.90)
Cause of CKD
Glomerular 146 0.86 (0.81 to 0.90)
0.05 (–0.02 to 0.13)
Nonglomerular 457 0.91 (0.85 to 0.95)
Error bars indicate 95% CI.
0.70 0.80 0.90 1.00 C statistics and 95% CIs closer to 1.00
C Statistic (95% CI) indicate better discrimination.
CKD indicates chronic kidney disease.

The results of the application of the KFRE to our cohort

Figure 2. Estimated vs Observed Probability of End-stage Renal Disease
are shown in Table 2. Overall, the equation provided excel- (ESRD) at 2 Years by Risk Group
lent discrimination and was similar whether the 4- or 8-vari-
able equation was used. As shown in Figure 1, although per- 50

formance of the KFRE was similar by race and cause of CKD, Estimated risk
40

Probability of ESRD, %
a statistically significant difference occurred in discrimina- Observed risk

tory performance by ethnicity and age. The C statistic was 30

higher in the Hispanic (0.96; 95% CI, 0.91-0.98) compared with
the non-Hispanic (0.86; 95% CI, 0.81-0.90) populations and 20
higher for those younger than 12 years (0.94; 95% CI, 0.91-
0.97) compared with those 12 years and older (0.85; 95% CI, 10

0.79-0.90). Of note, the percentage of Hispanic vs non-

Hispanic children who progressed to ESRD by 2 years was simi-
lar (9 [9.3%] vs 52 [10.4%], respectively), whereas 15 children
(5%) younger than 12 years progressed to ESRD by 2 years com-
pared with 47 (15.2%) of those 12 years or older.
Results from our model calibration are shown in Figure 2.
The cohort was divided into risk group tertiles based on the
2-year, 8-variable KFRE score as follows: estimated risk group
1 corresponds to a 2-year KFRE score of less than 2.6%; esti-
mated risk group 2, a KFRE score of at least 2.6% but less
than 13.2%; and estimated risk group 3, a KFRE score of at
least 13.2%. The median KFRE score for those in the highest
tertile was 33%. As depicted in Figure 2, the estimated pro-
portion of participants who developed ESRD closely matched
the observed proportion who progressed to ESRD at 2 years.
We compared the characteristics of children with a 2-year
KFRE score in the top tertile (>13%) with those of children in
the lower tertiles (≤13%) (Table 3). Overall, the median 2-year
KFRE score was 6% (IQR, 2%-21%). Of the entire cohort, 202
children (33.5%) had a 2-year KFRE score of greater than 13%,
and 55 (27.2%) of this group progressed to ESRD within 2 years
of follow-up compared with only 7 of 401 (1.7%) with a 2-year
KFRE score of 13% or less. Participants with a 2-year KFRE score
above 13% were notably older and more often male and had
higher blood pressures and lower hemoglobin levels at the time
of entry into our study.
One hundred forty-six participants (24.2%) had measured
ACR data available. For this subset of patients, no difference was
found in the 2-year risk discrimination using the KFRE (C sta-
0
1 2 3
Risk Group
Data represent the median estimated Kidney Failure Risk Equation (KFRE)
scores in each tertile of ESRD risk and the actual percentage of the cohort
(Kaplan-Meier estimate) who developed ESRD at 2 years. Estimated risk group 1
corresponds to participants with a 2-year KFRE score of less than 2.6%; risk
group 2, KFRE scores of at least 2.6% but less than 13.2%; and risk group 3,
KFRE risk score of at least 13.2%.
tistic, 0.96 for models including measured or calculated ACR;
95% CI for the difference in C statistics, −0.001 to 0.30).
Discussion
Few tools are currently available to estimate whether a child
with CKD will develop ESRD within 5 years. In this study, we
used a validated risk equation originally developed in an adult
cohort to determine whether equal risk discrimination could
be provided with the use of the KFRE in children. Overall, we
found that the KFRE score provides excellent discrimination
and calibration of the risk of ESRD in a pediatric cohort with
stages 3 and 4 CKD. The KFRE is a simple tool that uses rou-
tinely collected demographic and laboratory data and can eas-
ily be applied by all clinicians caring for children with CKD to
help provide anticipatory guidance, gauge timing of referral
to subspecialty care and transplant centers, and plan for the
optimal timing of dialysis access and kidney transplant.

jamapediatrics.com (Reprinted) JAMA Pediatrics February 2018 Volume 172, Number 2 177

© 2017 American Medical Association. All rights reserved.

Downloaded From: by a Sam Ratulangi University Hospital User on 04/24/2018

 Research Original Investigation Kidney Failure Risk Equation to Determine Risk of End-stage Renal Disease

Table 3. Comparison of Baseline Characteristics by 2-Year KFRE Score

KFRE Score
≤13% >13%
Characteristic (n = 401) (n = 202) P Valuea
Demographic
Age, median (IQR), y 11 (8-15) 12 (9-15) .71
Male, No. (%) 237 (59.1) 141 (69.8) .01
White race, No. (%) 270 (67.3) 131 (64.9) .54
Hispanic ethnicity, No. (%) 61 (15.2) 36 (17.8) .35
Glomerular cause of CKD, No. (%) 93 (23.2) 53 (26.2) .41
Physical examination
Systolic blood pressure
Mean (SD), mm Hg 107 (12) 110 (15) .002
Percentile based on age, sex, and height, 62.1 (33.0-83.1) 70.9 (42.6-91.6) .002
median (IQR)
Diastolic blood pressure
Mean (SD), mm Hg 65 (10) 69 (12) <.001
Percentile based on age, sex, and height, 64.0 (43.1-86.0) 74.4 (47.3-91.6) <.001
median (IQR)
Laboratory
eGFR, median (IQR), mL/min/1.73 m2 50 (44-56) 30 (25-34) <.001
Hemoglobin level, mean (SD), g/dL 13.0 (4.0) 11.9 (1.6) <.001 Abbreviations: CKD, chronic kidney
disease; eGFR, estimated glomerular
Serum calcium level, mean (SD), mg/dL 9.6 (0.5) 9.3 (0.8) <.001 filtration rate; ESRD, end-stage renal
Serum phosphorus level, mean (SD), mg/dL 4.3 (0.7) 4.9 (0.8) <.001 disease; IQR, interquartile range.
Serum albumin level, mean (SD), g/dL 4.4 (0.4) 4.0 (0.6) <.001 SI conversion factors: To convert
Serum bicarbonate level, mean (SD), mEq/L 21.7 (3.1) 19.5 (3.3) <.001 albumin to grams per liter, multiply by
10; bicarbonate to millimoles per liter,
Ratio of urine protein to creatinine levels, 283 (121-733) 1302 (567-2632) <.001 multiply by 1.0; calcium to millimoles
median (IQR), mg/g
per liter, multiply by 0.25; and
Ratio of urine albumin to creatinine level 131 (52-346) 565 (233-1179) <.001 phosphorus to millimoles per liter,
(calculated), median (IQR)
multiply by 0.323.
Outcome a
Calculated using the unpaired
Progression to ESRD within 2 y 7 (1.7) 55 (27.2) <.001 2-tailed t test, χ2 test, or Wilcoxon
from baseline visit, No. (%)
rank sum test as appropriate.

Although Tangri et al1 initially developed and validated the
KFRE in a large Canadian cohort, the KFRE was subsequently
validated in a cohort of European patients with CKD2 and was
found to accurately estimate risk of kidney failure, with the
best performance by the 8-variable equation. Performance of
the equation has also been validated in the African American
Study of Kidney Disease and Hypertension cohort,3 in which
use of the 1-year 4-variable equation improved ESRD risk es-
timation better than eGFR alone. In addition, Tangri et al4
assessed the accuracy of the KFREs for estimating the risk of
kidney failure using the CKD Prognosis Consortium with par-
ticipants from 30 countries and generally found that the equa-
tions remained highly accurate, although a calibration factor
was suggested for non–North American cohorts for which the
original equation overestimated risk. Our study is novel in the
application of this risk equation for the first time, to our knowl-
edge, to a large cohort of children with CKD.
Although an equation developed in adults who have dif-
ferent causes of their CKD performed well in children, the vari-
ables used in the KFRE notably include those that have been
previously identified as risk factors for CKD progression in
adults and children.5 Warady et al5 previously found that older
age, male sex, proteinuria, hypoalbuminemia, and hyperphos-
phatemia were associated with a shorter time to CKD progres-
sion in children with a nonglomerular cause of ESRD,
although for those with a glomerular cause, age, sex, and
hyperphosphatemia were not associated with ESRD risk. For
all causes of CKD, additional risk factors not accounted for in
the KFRE but previously identified as being important in chil-
dren included elevated blood pressure and anemia.5 In con-
trast to what would be expected based on the previously cited
study, a glomerular vs nonglomerular cause of CKD did not
significantly affect performance of the KFRE, although the
number of patients with glomerular causes of CKD was lower
in the CKiD cohort.5 In addition, those with glomerular causes
of CKD would be expected to have more albuminuria,21 which
is likely the main driver of CKD progression in this popula-
tion, as evidenced by the high C statistic with use of the KFRE
even among this subpopulation.
We found better performance of the KFRE in younger chil-
dren and those of Hispanic ethnicity. The reasons for differ-
ences by age are unclear, although we speculate that in younger
children, kidney disease tends to have nonglomerular causes
(congenital anomalies of the kidney and urinary tract) and
therefore may have a more predictable pattern of progres-
sion. In addition, progression of CKD in older children may be
subject to additional confounders, including nonadherence to
therapy22 and increased heterogeneity in the causes of CKD,
which may decrease the KFRE’s ability to discriminate the risk
of ESRD. We believe that the difference in performance of the

178 JAMA Pediatrics February 2018 Volume 172, Number 2 (Reprinted) jamapediatrics.com

© 2017 American Medical Association. All rights reserved.

Downloaded From: by a Sam Ratulangi University Hospital User on 04/24/2018

 Kidney Failure Risk Equation to Determine Risk of End-stage Renal Disease Original Investigation Research

KFRE by ethnicity deserves further exploration and valida-
tion, and we cannot definitively rule out the possibility that
this finding occurred by chance. We did not, on the other hand,
find statistically significant differences in performance of the
KFRE by race or cause of CKD.
Use of the KFRE may provide opportunities to improve the
care of children with CKD. The provision of better tools to es-
timate the timing of ESRD onset could allow for better timing
of living donor workup and planning for preemptive trans-
plant. Preemptive transplant is associated with better renal al-
lograft and patient survival among children compared with any
dialysis exposure9 and for this reason should be the preferred
treatment for children with ESRD. Preemptive transplant rates
could improve with better estimation of the timing of ESRD
onset, and future studies might evaluate the KFRE score that
may warrant transplant waitlist activation. For example, in the
adult literature, a 1-year KFRE score of greater than 10% should
prompt planning for ESRD care.3 Use of the KFRE in children
with CKD would not only be useful for pediatric nephrolo-
gists but also would have broader applicability to general
practitioners and adult nephrologists caring for children with
CKD. A number of states have no pediatric nephrologists16;
thus, the KFRE score would be useful in appropriately timing
referral to pediatric dialysis and transplant centers for end-
stage care. This use is especially important given that late re-
ferral to pediatric nephrology care has been associated with a
decreased likelihood of preemptive transplant.23 Vaccination
rates before kidney transplant may also be improved15 with bet-
ter estimation of the amount of time before ESRD onset. Spe-
cifically, an accelerated vaccine schedule might be appropri-
ate for the child rapidly progressing to ESRD, and revaccination
for children without protective antibody titers should be
strongly considered before kidney transplant, after which live
vaccines are contraindicated.14 In addition, all clinicians could
use the information gained from use of the KFRE to provide
counseling and guide families in the planning of important life
decisions regarding finances, housing, and education. How-
ever, we would emphasize that a low 5-year KFRE score does
not mean that the lifetime risk of ESRD is necessarily low. Given
the long-term survival of most pediatric patients with CKD, the
lifetime risk of kidney failure may still be significant, and on-
going interventions that retard the progression of CKD should
still be aggressively pursued.
Strengths and Limitations
Our study has several strengths and limitations. Our study is
strengthened by the relatively large size of this pediatric CKD
cohort with national representation and comprehensive
follow-up data, including ascertainment of time to ESRD. Our
study is limited by missing albuminuria measurements, al-
though we were able to estimate albuminuria from protein-
uria measurements using a previously validated approach with
high correlation.3 Our study is also limited by the median fol-
low-up time of only 3.8 years, which may explain the rela-
tively lower C statistic for the 5-year KFRE score compared with
shorter-term risk scores in this cohort. Our study results may
not be generalizable to the larger population of children with
CKD, given that enrollment in the CKiD study is voluntary and
therefore may select for more adherent patients and families.
Conclusions
We found that use of the KFRE in children with CKD accu-
rately discriminates risk of ESRD. We propose that this equa-
tion be incorporated into the clinical care of children with CKD
to aid clinicians, patients, and their families in planning for end-
stage care.

ARTICLE INFORMATION Institute of Diabetes and Digestive and Kidney REFERENCES

Accepted for Publication: September 18, 2017.
Published Online: December 18, 2017.
doi:10.1001/jamapediatrics.2017.4083
Author Contributions: Drs Winnicki and Ku had full
access to all the data in the study and take
responsibility for the integrity of the data and the
accuracy of the data analysis.
Study concept and design: Winnicki, Furth, Ku.
Acquisition, analysis, or interpretation of data: All
authors.
Drafting of the manuscript: Winnicki, Ku.
Critical revision of the manuscript for important
intellectual content: All authors.
Statistical analysis: Winnicki, McCulloch.
Obtained funding: Furth, Warady, Ku.
Administrative, technical, or material support:
Warady.
Study supervision: Furth, Ku.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported by
grants HL131023 (Dr Ku) and DK090070
(Dr Mitsnefes) from the National Institutes of
Health. The Chronic Kidney Disease in Children
Cohort Study (CKiD) was conducted by the CKiD
investigators and supported by the National
Diseases (NIDDK), with additional support from
grants U01-DK-66143, U01-DK-66174, and
U01-DK-66116 from the National Institute of Child
Health and Human Development and the National
Heart, Lung, and Blood Institute, and NIDDK and
grants U01DK-082194 and U01-DK-66116 from the
NIDDK. The data and samples from the CKiD study
reported herein were supplied by the NIDDK
Central Repositories.
Role of the Funder/Sponsor: The sponsors had no
role in the design and conduct of the study;
collection, management, analysis, and
interpretation of the data; preparation, review, or
approval of the manuscript; and decision to submit
the manuscript for publication.
Disclaimer: This article does not necessarily reflect
the opinions or views of the CKiD study, the NIDDK
Central Repositories, or the NIDDK.
Additional Contributions: Barbara Grimes, PhD,
University of California, San Francisco, contributed
to and verified the statistical analysis. She was
compensated for this work.
1. Tangri N, Stevens LA, Griffith J, et al. A predictive
model for progression of chronic kidney disease to
kidney failure. JAMA. 2011;305(15):1553-1559.
2. Peeters MJ, van Zuilen AD, van den Brand JA,
Bots ML, Blankestijn PJ, Wetzels JF; MASTERPLAN
Study Group. Validation of the kidney failure risk
equation in European CKD patients. Nephrol Dial
Transplant. 2013;28(7):1773-1779.
3. Grams ME, Li L, Greene TH, et al. Estimating time
to ESRD using kidney failure risk equations: results
from the African American Study of Kidney Disease
and Hypertension (AASK). Am J Kidney Dis. 2015;65
(3):394-402.
4. Tangri N, Grams ME, Levey AS, et al; CKD
Prognosis Consortium. Multinational assessment of
accuracy of equations for predicting risk of kidney
failure: a meta-analysis. JAMA. 2016;315(2):164-174.
5. Warady BA, Abraham AG, Schwartz GJ, et al.
Predictors of rapid progression of glomerular and
nonglomerular kidney disease in children and
adolescents: the Chronic Kidney Disease in Children
(CKiD) cohort. Am J Kidney Dis. 2015;65(6):878-888.
6. Mange KC, Joffe MM, Feldman HI. Effect of the
use or nonuse of long-term dialysis on the

jamapediatrics.com (Reprinted) JAMA Pediatrics February 2018 Volume 172, Number 2 179

© 2017 American Medical Association. All rights reserved.

Downloaded From: by a Sam Ratulangi University Hospital User on 04/24/2018

 Research Original Investigation
subsequent survival of renal transplants from living
donors. N Engl J Med. 2001;344(10):726-731.
7. Ellis EN, Martz K, Talley L, Ilyas M, Pennington
KL, Blaszak RT. Factors related to long-term renal
transplant function in children. Pediatr Nephrol.
2008;23(7):1149-1155.
8. Vats AN, Donaldson L, Fine RN, Chavers BM.
Pretransplant dialysis status and outcome of renal
transplantation in North American children:
a NAPRTCS Study: North American Pediatric Renal
Transplant Cooperative Study. Transplantation.
2000;69(7):1414-1419.
9. Amaral S, Sayed BA, Kutner N, Patzer RE.
Preemptive kidney transplantation is associated
with survival benefits among pediatric patients with
end-stage renal disease. Kidney Int. 2016;90(5):
1100-1108.
10. Groothoff JW. Long-term outcomes of children
with end-stage renal disease. Pediatr Nephrol.
2005;20(7):849-853.
11. Goldstein SL, Graham N, Burwinkle T, Warady B,
Farrah R, Varni JW. Health-related quality of life in
pediatric patients with ESRD. Pediatr Nephrol.
2006;21(6):846-850.
180 JAMA Pediatrics February 2018 Volume 172, Number 2 (Reprinted)
© 2017 American Medical Association. All rights reserved.
Downloaded From: by a Sam Ratulangi University Hospital User on 04/24/2018
Kidney Failure Risk Equation to Determine Risk of End-stage Renal Disease
12. Rana A, Gruessner A, Agopian VG, et al. Survival
benefit of solid-organ transplant in the United
States. JAMA Surg. 2015;150(3):252-259.
13. Jung HW, Kim HY, Lee YA, et al. Factors
affecting growth and final adult height after
pediatric renal transplantation. Transplant Proc.
2013;45(1):108-114.
14. Neu AM. Immunizations in children with
chronic kidney disease. Pediatr Nephrol. 2012;27
(8):1257-1263.
15. Lee DH, Boyle SM, Malat G, Sharma A, Bias T,
Doyle AM. Low rates of vaccination in listed kidney
transplant candidates. Transpl Infect Dis. 2016;18(1):
155-159.
16. Ku E, Johansen KL, Portale AA, Grimes B, Hsu
CY. State level variations in nephrology workforce
and timing and incidence of dialysis in the United
States among children and adults: a retrospective
cohort study. BMC Nephrol. 2015;16:2-11.
17. National Institute of Diabetes and Digestive and
Kidney Diseases. NIDDK Central Repository.
https://www.niddkrepository.org/studies/ckid/.
Updated August 22, 2017. Accessed February 21,
2015.
18. Furth SL, Cole SR, Moxey-Mims M, et al. Design
and methods of the Chronic Kidney Disease in
Children (CKiD) prospective cohort study. Clin J Am
Soc Nephrol. 2006;1(5):1006-1015.
19. Furth SL, Abraham AG, Jerry-Fluker J, et al.
Metabolic abnormalities, cardiovascular disease risk
factors, and GFR decline in children with chronic
kidney disease. Clin J Am Soc Nephrol. 2011;6(9):
2132-2140.
20. Schwartz GJ, Muñoz A, Schneider MF, et al.
New equations to estimate GFR in children with
CKD. J Am Soc Nephrol. 2009;20(3):629-637.
21. Wong CS, Pierce CB, Cole SR, et al; CKiD
Investigators. Association of proteinuria with race,
cause of chronic kidney disease, and glomerular
filtration rate in the Chronic Kidney Disease in
Children study. Clin J Am Soc Nephrol. 2009;4(4):
812-819.
22. Beier UH, Green C, Meyers KE. Caring for
adolescent renal patients. Kidney Int. 2010;77(4):
285-291.
23. Boehm M, Winkelmayer WC, Arbeiter K,
Mueller T, Aufricht C. Late referral to paediatric
renal failure service impairs access to pre-emptive
kidney transplantation in children. Arch Dis Child.
2010;95(8):634-638.
jamapediatrics.com