Pediatr Nephrol (2010) 25:16991706 DOI 10.

1007/s00467-010-1538-6

ORIGINAL ARTICLE

Risk for anemia in pediatric chronic kidney disease patients: a report of NAPRTCS
Meredith A. Atkinson & Karen Martz & Bradley A. Warady & Alicia M. Neu

Received: 11 June 2009 / Revised: 26 March 2010 / Accepted: 30 March 2010 / Published online: 13 May 2010 # IPNA 2010

Abstract Previous studies in children with chronic kidney disease (CKD) have identified low hemoglobin as a risk factor for poor outcomes. A retrospective review of the North American Pediatric Renal Trials and Collaborative Studies (NAPRTCS) CKD registry was performed to identify the prevalence of and risk factors for anemia among children with stage 3 CKD, including both patients with low hemoglobin and those whose hemoglobin normalized with an erythropoiesis-stimulating agent (ESA). At enrollment, 2,792 patients had stage 3 CKD. Mean age was 9.5 (0.11) years, 62.1% were male, 61.3% were white, and 43.7% had structural/urologic disease. Among 1,640 of those patients with 12 month follow-up data available for multivariate analysis, 73% met the criteria for anemia. Multivariate logistic regression analysis identifying risk factors for anemia at the 12-month follow-up revealed that, after controlling for estimated glomerular filtration rate, age >2 years, male sex, earlier era of study entry, and prescription of anti-hypertensive medications are associated with an increased risk for anemia at 12 months. In addition, multivariate Cox proportional hazards regression analysis
The views expressed in this manuscript are those of the authors and do not necessarily reflect the official policy of NAPRTCS. M. A. Atkinson (*) : A. M. Neu Division of Pediatric Nephrology, Johns Hopkins University, 200 N. Wolfe St, Baltimore, MD 21287, USA e-mail: matkins3@jhmi.edu K. Martz EMMES Corporation, Rockville, MD, USA B. A. Warady Section of Pediatric Nephrology, Childrens Mercy Hospital, Kansas City, MO, USA

revealed that when patients with ESA-corrected hemoglobin are included in the definition, anemia is not associated with increased risk of progression to end stage renal disease (dialysis initiation or transplantation). Keywords Anti-hypertensives . CKD stage . Erythropoietin stimulating agent . Progression . Risk factors

Introduction Low hemoglobin is a common co-morbidity in children with chronic kidney disease (CKD) and is associated with multiple adverse clinical consequences, including the development and progression of cardiovascular risk factors such as left ventricular hypertrophy [1, 2]. In addition, higher hemoglobin values in children with CKD have been associated with better health-related quality of life [3], and there is data in the literature to suggest that uncontrolled anemia may be related to CKD progression [46]. A recent study done in pediatric CKD patients found an increased risk of hospitalization in children with low hemoglobin compared to those with normal hemoglobin [7]. Thus, anemia management is a core component of clinical nephrology practice. Although many factors, including shortened red blood cell life span and iron deficiency, contribute to CKD-associated anemia, erythropoietin deficiency is the primary etiologic factor. Recent studies in pediatric CKD patients have sought to identify the prevalence of patients with low hemoglobin or hematocrit values and to evaluate for associations between this unrecognized or undertreated anemia and outcomes, excluding patients whose erythropoietin deficiency has been effectively treated with an erythropoiesis-stimulating agent (ESA) [7, 8]. In contrast, there have been no studies to

1700

Pediatr Nephrol (2010) 25:16991706

identify the prevalence and the potential risk factors for and consequences of anemia in pediatric CKD patients, including those children in whom hemoglobin or hematocrit values are normalized on ESA treatment. Including patients with ESA-corrected hemoglobin in the definition of anemia is important in order to further clarify the impact of anemia on outcomes in children with CKD. For example, is the association of anemia and risk for CKD progression related to low hemoglobin alone, or is the development of CKDassociated anemia a reflection of lower kidney function beyond that detected by estimates of glomerular filtration rate (GFR)? We therefore sought to use a large, national database of children with CKD, the North American Pediatric Renal Trials and Collaborative Study (NAPRTCS) CKD registry, to examine the prevalence of anemia, independent of correction of the hemoglobin value, through a retrospective cohort study. Additional aims were to identify risk factors for anemia at 12 months of follow-up among children enrolled in this registry and to evaluate for any potential association between anemia and risk for progression of CKD.

Materials and methods We conducted a retrospective review of the NAPRTCS CKD Registry. NAPRTCS is a voluntary, collaborative research effort of more than 100 pediatric centers in Canada, Mexico, and the USA that collects data on children who have CKD, are on dialysis, and have undergone kidney transplantation. Data included in the registry are deidentified. Participation in the NAPRTCS registry is approved by the institutional review board at each center. Data on children with CKD have been collected by NAPRTCS since 1994 and as of the 2008 Annual Report, the registry contains information on more than 7000 patients with CKD as defined by an estimated GFR by the Schwartz formula of 75 ml/min/1.73 m2 [9]. Data are collected at study enrollment and then every 6 months until an endpoint (kidney transplantation, initiation of dialysis, loss to follow-up, or death) is reached. Anemia was defined as treatment with an ESA or hemoglobin <5th percentile for children 2 years of age or >2 standard deviations (SD) below the mean for children 024 months of age. These hemoglobin cutoffs are based on age- and sex-specific values below which an anemia work-up is suggested, as outlined in the 2006 Kidney Disease Outcomes Quality Initiative (KDOQI) Clinical Practice Guidelines and Clinical Practice Recommendations for Anemia in Chronic Kidney Disease [10]. As over 80% of patients in the NAPRTCS registry have hematocrit values, but not hemoglobin values recorded (hemoglobin has been included on the data collection forms only in recent years), we

calculated hemoglobin values for the latter patients by multiplying the hematocrit values by 0.3 [10]. Using a cross-sectional analysis design, we explored the prevalence of anemia by stage of CKD at study registration, using the KDOQIs CKD staging guidelines [11]. For the multivariate analysis to identify risk factors for anemia at 12 months of follow-up, we restricted our study to the subcohort of children with stage 3 CKD (as defined by an estimated GFR of 30 to <60 ml/min/1.73 m2) at registration; this was a convenient sample with a significant number of patients with 12-month follow-up data available, at a stage of CKD at which anemia becomes more prevalent [12]. The data studied included age, sex, and cause of CKD (structural/urologic, focal segmental glomerulosclerosis, and others). These diagnostic categories were used because they represent the leading causes of CKD in children. Other causes of CKD, including chronic glomerulonephritis and polycystic kidney disease, are present in <3% of patients in the registry overall and are therefore evaluated together as other. Race was also included, which in NAPRTCS is collected as either white, black, Hispanic, or other. Other baseline variables included in the analyses were standardized height and weight, year of study registration, and prescription of recombinant human growth hormone, ESAs, antihypertensive medications, and vitamin D. Height and weight standard deviation scores (SDS) were calculated using the age- and sex-specific formulae provided by the Centers for Disease Control, giving normative values for USA children. [13] A height SDS of less than 1.88 is equivalent to less than the third percentile. Parathyroid hormone (PTH) data are included in the NAPRTCS database, but are recorded as either less than or more than twice the upper normal limit. Due to a significant amount of missing data, PTH was not included as a variable in the multiple logistic regression portion of this analysis. Clinical indicators of iron status, including serum ferritin and transferrin saturation, are not collected by NAPRTCS and thus could not be evaluated as covariates in our analysis. To be included in the multiple logistic regression analysis, children with stage 3 CKD at baseline had to have 12-month follow up data available, including hematocrit or hemoglobin and ESA use, from which to determine anemia status. Statistical methods Descriptive statistics were used to compare anemic versus non-anemic patients in CKD stage 3 at registration. Data are reported as percentages or means standard error (SE), as appropriate. Frequency distributions were compared using the Pearsons chi-square test. Continuous factors were compared using the non-parametric Wilcoxon test.

Pediatr Nephrol (2010) 25:16991706

1701
100
Hemoglobin in anemia range for age & sex Anemia (hgb in anemia range or ESA) Treatment Failure (hgb in anemia range & ESA)

Multivariate logistic regression analysis was used to identify factors associated with anemia at 12 months postregistration. We considered the following variables collected at registration: age, race, sex, primary diagnosis, year of registration, use of erythropoietin, use of human growth hormone, use of antihypertensive medications, use of vitamin D supplements, and height and weight SDS. These factors were entered into the model and removed one at a time, the least significant first, until all remaining factors were significant at the 0.05 level. A Cox proportional hazards regression analysis was performed to characterize the association between anemia and the development of end-stage renal disease (ESRD). Progression to ESRD was defined as the termination of CKD status due to either dialysis initiation or transplant. This analysis included anemia status at registration as a co-variate. Other covariates included in the model were age, race sex, primary diagnosis, creatinine clearance, albumin, phosphorus, calcium and blood urea nitrogen (BUN) levels. Model reduction was performed using the backward elimination method, and factors remaining significant at p value 0.05 were maintained in the final model. All statistical analyses were performed using the SAS System for Windows ver. 9.02 (SAS Institute, Cary, NC).

80

Percent

60

40

20

0
Stage 2 (n=940) Stage 3 (n=2792) Stage 4 (n=1742) Stage 5 (n=792)

Fig. 1 Anemia prevalence at baseline in the North American Pediatric Renal Trials and Collaborative Study (NAPRTCS) chronic kidney disease (CKD) Registry (n=6242). ESA Erythropoiesis-stimulating agent, hgb hemoglobin

Results Figure 1 presents anemia prevalence using the age- and sexspecific parameters by CKD stage at baseline in the NAPRTCS CKD registry among the 6242 children for whom anemia data were available. The figure clearly demonstrates that the prevalence of both low hemoglobin and anemia defined as low hemoglobin or ESA use regardless of hemoglobin value increase as CKD stage advances. Approximately 73% of children at CKD stage 3, 87% at CKD stage 4, and over 93% at CKD stage 5 were defined as anemic using our definition. At the time of enrollment in the NAPRTCS, the prevalence of children prescribed an ESA who continue to demonstrate low hemoglobin values is fairly low at earlier CKD stages (7% at stage 3) but it is higher among patients with more advanced CKD and is >40% among patients with CDK stage 5. In total, 2792 patients met the criteria for stage 3 CKD at registration. The demographic and clinical characteristics of this cohort are summarized in Table 1. Mean age was 9.5 years. As is fairly typical of pediatric CKD, the majority of patients were noted to be male (62%) and white (61%). Approximately 73% of patients were noted to be anemic at registration. A significant proportion, 43.7%, was noted to have structural/urologic disease (obstructive uropathy, reflux nephropathy, or renal dysplasia) as the underlying

cause of CKD. Nearly one-third of patients had evidence of significant growth failure with height SDS more negative than 1.88. Over 46% of children were missing PTH data. Nearly 11% were prescribed an ESA, 6% were prescribed recombinant human growth hormone at registration, and 44% were prescribed antihypertensive medication. Of the 2792 stage 3 CKD patients, the anemia status of 1640 children was available at 12 months of follow-up with complete co-variate data and was included in the multivariate analysis. Based on our definition at 12 months, 73% of children were defined as being anemic. Demographic and clinical characteristics of this cohort at 12 months by anemia status are given in Table 2. There was a significant difference in mean age between groups (10.3 years in anemic patients vs. 6.8 years in non-anemic patients, p< 0.001), and there were fewer anemic patients among those 02 years of age at registration and more anemic patients in the 6- to 12-year-old and 13-year-old age groups. There was no difference in either sex or racial distribution between groups. Anemic patients compared to those without anemia were more likely to be prescribed antihypertensive medications and vitamin D, but there was no difference in the proportions of patients prescribed growth hormone. Information about particular antihypertensive or other medication subclasses or preparations within the NAPRTCS CKD registry data was not available. The anemic patients were less likely to have structural/urologic disease as the primary cause of CKD. Patients with anemia had higher mean weight SDS compared to non-anemic patients (0.66 vs. 1.03, p=0.003), but there was no difference in mean height SDS. A higher proportion of anemic patients were noted to have PTH values greater than

1702 Table 1 Patient characteristics at registration in children with stage 3 chronic kidney disease (n=2792) Characteristic Age, years, mean (SE) Male, % (n) Race, % (n) White Black Hispanic Other Anemic, % (n)a Hemoglobin, g/dL, mean (SE) Primary diagnosis, % (n) Structural/Urologic FSGS Other Missing Height Z-score, mean (SE) Weight Z-score, mean (SE) Height SDS 1.88, % (n) Parathyroid hormone, % (n) <2 Upper normal limit >2 Upper normal limit Missing Prescribed ESA, % (n) Prescribed anti-hypertensives, % (n) Prescribed growth hormone, % (n) Prescribed vitamin D, % (n) Year of registration, % (n) 19941995 19961998 19992001 20022004 20052007 Values 9.5 (0.11) 62.1 (1734) 61.3 (1712) 18.0 (504) 14.0 (391) 6.5 (182) 72.5 (2024) 10.8 (0.03) 43.7 (1221) 9.3 (259) 45.1 (1259) 1.9 (53) -1.31 (0.03) -0.69 (0.03) 32.6 (910) 38.4 (1073) 14.9 (415) 46.7 (1304) 10.5 (294) 44.2 (1235) 5.8 (163) 31.1 (869) 30.9 27.6 17.2 13.8 10.5 (863) (770) (480) (385) (294)

Pediatr Nephrol (2010) 25:16991706

FSGS, Focal segmental glomerulosclerosis; SDS, standard deviation score; ESA, erythropoiesis-stimulating agent Data are presented as the percentage with the number in parenthesis, or as the mean with the standard error (SE) in parenthesis, whichever is the more appropriate
a

Anemia is defined as hemoglobin in the anemia range for age and sex or treatment with ESA

twofold the upper limit of normal compared to non-anemic patients, but both groups were also noted to have a significant proportion of missing data. Year of study entry was significantly different between groups, with higher proportions of anemic patients registered in earlier years of the NAPRTCS CKD cohort study. Finally, those children with anemia at 12 months had significantly lower estimated GFR than those without anemia (43.3 vs. 45.5 ml/min/ 1.73 m2, p<0.001).

Table 3 gives the results of the multivariate logistic regression analysis used to identify risk factors for anemia at 12 months of follow-up in children with stage 3 CKD. Age group at baseline was found to be a significant predictor of anemia at follow-up, with those children >2 years of age more likely to be anemic at 12 months compared to those <2 years [25 years, odds ratio (OR) 4.48, 95% confidence interval (CI) 2.896.95; 612 years, OR 16.43, 95% CI 10.6125.45; 13 years, OR 16.40, 95% CI 10.3226.08]. Female patients were noted to be less likely to demonstrate anemia at age 1 year than their male counterparts (OR 0.69, 95% CI 0.520.93). Primary cause of CKD was not a significant predictor of anemia at 12 months. Year of registration in the NAPRTCS cohort was a significant predictor, with those children enrolled in 1999 and later less likely to be anemic than those enrolled from 19941995 (19992001, OR 0.56, 95% CI 0.360.88; 20022004, OR 0.18, 95% CI 0.120.28; 20052007, OR 0.25, 95% CI 0.150.42). Children enrolled between 1996 and 1998 had no significant difference in risk compared to the 19941995 cohort. Children prescribed antihypertensive medications at baseline were more likely to be anemic at 12 months than those not prescribed these medications (OR 1.66, 95% CI 1.222.25). Finally, increasing eGFR (per 1 ml/min/1.73 m2 increase) was associated with a small but significantly decreased risk of anemia during follow-up (OR 0.96, 95% CI 0.950.98). Baseline values of hematocrit and use of ESA were also included in the model and were, as expected, highly predictive of risk of anemia at 12 months. Factors removed from the model because they were not predictive of anemia at 12 months of follow-up included race, the use of vitamin D, the use of growth hormone, and height and weight SDS. In addition, in a subanalysis of 907 children for whom PTH data were available, PTH was not significantly associated with the risk of anemia at 12 months (p=0.1030) when included as a variable in the multivariate model as less than versus double the upper limit of normal. A multivariate Cox proportional hazards regression analysis was performed within this population of children with stage 3 CKD to characterize the association between anemia and risk for progression to ESRD. Baseline factors, including anemia status, sex, race, age, primary diagnosis, creatinine clearance, albumin, phosphate, calcium, and BUN levels, were included in the initial model. Anemia status was not found to be a significant risk factor for progression of CKD.

Discussion Defining normal hemoglobin values and anemia in children with CKD has been controversial, and the definition has

Pediatr Nephrol (2010) 25:16991706 Table 2 Patient characteristics at registration in children with stage 3 chronic kidney disease by anemia status at 12 months of follow-up (n=1640) Characteristic Patients with anemia at 1 year (n=1195) 10.3 (0.14) 7.3 15.7 41.3 35.8 62.2 64.9 16.2 13.3 6.5 10.3 (0.05) Patients with no anemia at 1 year (n=445) 6.8 (0.29) 32.6 22.0 21.8 23.6 62.2 61.1 18.9 13.3 6.7 11.7 (0.09)

1703 p value <0.001 <0.001

Age, years, mean (SE) Age, years (%) 01 25 612 13+ Male (%) Race (%) White Black Hispanic Other Hemoglobin, g/dL, mean (SE) Primary diagnosis (%) Structural/Urologic FSGS Other Missing Height Z-score, mean (SE) Weight Z-score, mean (SE) Height SDS -1.88 (%) Parathyroid hormone (%) <2 Upper normal limit >2 Upper normal limit Missing Prescribed ESA (%) Prescribed antihypertensives (%) Prescribed growth hormone (%) Prescribed vitamin D (%) Year of registration (%) 19941995 19961998 19992001 20022004 20052007 eGFR (ml/min/1.73m2)

0.994 0.435

<0.001 <0.001

42.5 7.9 47.4 2.1 -1.32 (0.04) -0.66 (0.05) 33.1 38.5 18.6 42.9 12.7 47.4 6.8 37.2 40.4 28.9 15.9 9.5 5.4 43.3 (0.24)

55.5 6.5 36.0 2.0 -1.47 (0.08) -1.03 (0.09) 35.1 47.6 9.9 42.5 5.8 29.4 4.9 26.1 24.7 22.2 12.8 24.9 15.3 45.5 (0.39)

0.141 0.003 0.370 <0.001

<0.001 <0.001 0.173 <0.001 <0.001

eGFR, Estimated glomerular filtration rate Data are presented as the percentage or as the mean with the SE in parenthesis, whichever is the more appropriate
Anemia defined as hemoglobin in the anemia range for age and sex or treatment with ESA
a

<0.001

evolved in recent years. The 2002 KDOQI clinical practice guidelines defined anemia as hemoglobin values <11 g/ dL (hematocrit <33%), based in large part on evidence in adults that outcomes were worse with hemoglobin values below this threshold [6, 11]. However, with increasing recognition that normal hemoglobin values in healthy children vary by age and sex, the 2006 KDOQI guidelines for anemia in pediatric CKD used the Third National Health

and Nutrition Examination Survey (NHANES III) reference data for normative hemoglobin values in children and recommend initiation of evaluation for anemia once the hemoglobin value is less than the 5th percentile for age and sex or >2 SD below mean hemoglobin values in children 0 24 months of age [10, 14]. It has been noted that the clinical use of hemoglobin >11 g/dL/hematocrit <33% as a threshold likely underestimates the prevalence of anemia

1704 Table 3 Multivariate logistic regression to identify risk factors for anemia at the 12-month follow-up in children with stage 3 chronic kidney disease Factor Age at registration (<2 years referent) 25 years 612 years 13 years Sex (female vs. male referent) Year of registration (19941995 referent) 19961998 19992001 20022004 20052007 Antihypertensive medications (yes vs. no) eGFR (per 1 ml/min/1.73 m2 increase) Odds ratio

Pediatr Nephrol (2010) 25:16991706 95% confidence Interval p value

4.48 16.43 16.40 0.69 0.71 0.56 0.18 0.25 1.66 0.96

2.896.95 10.6125.45 10.3226.08 0.520.93 0.501.01 0.360.88 0.120.28 0.150.42 1.222.25 0.950.98

<0.001 <0.001 <0.001 0.013 0.058 0.011 <0.001 <0.001 0.001 <0.001

Model is adjusted for hemoglobin (g/dL) at registration and ESA use at registration

in the pediatric CKD population, while the use of an ageand sex-specific definition increases the number of patients who meet criteria for anemia overall [6, 7]. Our cross-sectional analysis has demonstrated that among the NAPRTCS CKD cohort, the prevalence of anemia, both corrected with ESA and uncorrected, increases as CKD stage advances. In addition, the proportion of patients demonstrating hemoglobin values below target range for age and sexdespite treatment with an ESAincreases at advancing CKD stages, reaching >40% by stage 5 (Fig. 1). The reasons for this uncorrected anemia may vary, and although further evaluation of the relationship between ESA dose and anemia is important, much of the data necessary to perform these analyses are either missing or unavailable in the NAPRTCS registries. Specifically, we are unable to assess for iron deficiency status within the NAPRTCS registry, as data on iron stores are not routinely collected. Some of the uncorrected anemia may be attributed to issues with inadequate ESA dosing, which is an increasingly recognized problem in the treatment of anemia among the pediatric CKD population, particularly in dialysis patients [15, 16]. Children on dialysis may require proportionally larger ESA doses than adults to achieve comparable hemoglobin values [15]. However, the data collected by NAPRTCS do not allow the accurate detection of inadequate dosing. Although ESA dose is collected at registration and every 6 months, the starting date for that dose and/or any recent changes to that dose are not collected. ESA doses may be changed frequently, and so a single ESA dose may not be reflective of true ESA exposure. Patient compliance with ESA therapy, which is largely a home therapy in children with CKD, may also play a role, and medication adherence is not collected by the NAPRTCS [14]. Finally, multiple other clinical factors that may contribute to a poor response to ESAs among patients with CKD, including decreased red blood cell survival and inflammation, cannot be assessed using the

registry data [1719]. Given these limitations, in this study we did not attempt to further elucidate the relationship between ESA dose and hemoglobin level. Accurate assessment of this question will likely require a prospective trial. However, this study does provide important information as it is the first to identify the prevalence of risk factors for, and the risks of anemia, even among those with corrected hemoglobin values on ESA treatment [7, 20]. In this analysis, we found the prevalence of anemia to consistently increase with advancing CKD stage, with a prevalence of >70% in children with stage 3 CKD at study registration. Furthermore, there is a high prevalence of anemia in children with stage 3 CKD not only at baseline, but also at 12 months of follow-up, with 73% of patients meeting anemia criteria. By including ESA use in our definition of anemia, we were able to identify not only those children with low hemoglobin values, but also those with treated anemia, regardless of the normalization of actual hemoglobin. Children with stage 3 CKD enrolled in the NAPRTCS registry for whom 12-month follow-up data have been collected can be assumed to have been regularly followed by a pediatric nephrologist for at least 1 year. It is therefore unlikely that issues such as poor access to care have a significant effect on the 12-month anemia prevalence we observed. In this study, we have identified risk factors for anemia among children specifically followed by pediatric nephrologists at pediatric centers. Studies of adults with CKD have identified multiple factors associated with increased anemia risk, including AfricanAmerican race, female sex, and advancing age (specifically age >75 years) [21, 22]. Associations between anemia and the primary cause of CKD have also been identified, with those adult patients with underlying diabetes and hypertension demonstrating the highest risk [21, 22]. In contrast to adult data, in our analysis, female sex was actually found to be associated with a decreased risk of anemia at 12 months; this is likely due to our use of sex-specific hemoglobin

Pediatr Nephrol (2010) 25:16991706

1705

values to define anemia within our cohort, which accounts for lower hemoglobin values in adolescent females. Race was not significantly associated with anemia risk at 12 months of follow-up in this analysis. However, our study is among the first to identify specific risk factors for continued/developing anemia in children at 12 months, which include age >2 years, era of entry into the NAPRTICS cohort, and treatment with antihypertensive medications at baseline. The decreased risk of anemia seen among children <2 years of age at baseline may be related to the relative increase in GFR that occurs physiologically in all children throughout the first 2 years of life: younger infants with estimated GFR values at baseline which placed them in the category of stage 3 CKD may in fact have undergone an improvement in GFR over the 12-months follow-up period related to physiological increases in single nephron GFR, and thus may have been less likely to have erythropoietindeficiency anemia at 12 months due to less advanced CKD. Year of registration in the NAPRTCS study was also found to be significantly associated with anemia risk. Compared to those children enrolled in the initial years of the NAPRTCS CKD cohort (19941995), those enrolled from 1999 onwards are noted to have decreased anemia risk at 12 months. This can likely be attributed to improvements in the clinical management of anemia in children with CKD in recent years, including an increasing awareness of the prevalence of anemia in this population as well as a more aggressive approach to the correction of factors that may contribute to anemia, including iron deficiency and hyperparathyroidism. In fact, the first CKD-specific (non-dialysis) guidelines for the diagnosis and management of anemia were released by KDOQI as recently as 2000 [10]. The prescription of antihypertensive medications was also found to be associated with a 66% increased risk of anemia at 12 months. The requirement for antihypertensive medications may be a marker of children with more severe kidney disease, who also may be more likely to be erythropoietin-deficient. Although angiotensin-converting enzyme inhibitors specifically have been associated with increased risk of anemia [23], the NAPRTCS database has no information available on the subclasses of antihypertensive medications prescribed, and thus we could not examine this effect within the anemic patients in our cohort. Uncontrolled hypertension is a known and independent risk factor for CKD progression in adults, and treatment with anti-hypertensive medications at baseline may serve to identify children with an increased risk of progression overall, which in turn may be related to persistent or worsening erythropoietin deficiency and anemia [5]. In the univariate analysis, anemic patients were more likely to be prescribed vitamin D than non-anemic patients, and the need for treatment with these compounds could also be

construed to indicate more severe/advanced disease. However, the use of vitamin D failed to significantly predict anemia at 12 months of follow-up in the multivariate analysis. In addition, although nearly half of children in the multivariate analysis were missing PTH data, a sub-analysis among those children for whom PTH data had been recorded failed to identify an association between PTH level and anemia risk. The underlying cause of kidney disease was not significantly associated with 12-month anemia risk in this cohort. However, an association between change in GFR and anemia risk was noted. Even among children with preexisting stage 3 CKD, small, incremental (by 1 ml/min/ 1.73 m2) increases in the estimated GFR were associated with a small but significant decrease in anemia risk. This association suggests that clinical strategies and interventions implemented to slow GFR decline in children with CKD may be effective not only in delaying the onset of ESRD but also for minimizing the burden of comorbidities, such as anemia. The multivariate analysis revealed, after controlling for other factors, such as sex, age, race, primary diagnosis, and estimated GFR, that anemia was not associated with risk of progression to ESRD among children with stage 3 CKD in this cohort. While other studies have suggested that low hemoglobin may be a risk factor for CKD progression [46], by including patients in whom hemoglobin is normalized with an ESA, our analysis suggests that the development of anemia per se may not predict progression. It should be noted that our use of age- and sex-specific hemoglobin values to define anemia in our analysis may make our results less comparable to those of previous studies. There are additional limitations to this study. In this retrospective analysis, inferences about causality cannot be made. In addition, although the NAPRTCS CKD registry is among the largest pediatric CKD databases available, it is a voluntary registry; thus, the enrolled patients may not be representative of all North American children with CKD but rather of those seen at tertiary care centers with active pediatric nephrology programs. Although hyperparathyroidism is a known risk factor for the anemia of CKD, due to missing data we could not fully evaluate the association of hyperparathyroidism with risk of anemia at 12 months in the full cohort [17]. However, the use of vitamin D compounds, which may represent a proxy for clinical hyperparathyroidism, was not a significant predictor of anemia at 12 months in this analysis. We were not able to assess the contribution of iron deficiency to risk for anemia in this cohort due to the lack of data on clinical markers of iron stores within the NAPRTCS CKD registry. In conclusion, using age- and sex-specific hemoglobin values to define anemia, we have confirmed the increasing

1706

Pediatr Nephrol (2010) 25:16991706 9. Schwartz GJ, Brion LP, Spitzer A (1987) The use of plasma creatinine concentration for estimating glomerular filtration rate in infants, children, and adolescents. Pediatr Clin North Am 34 (3):571590 10. KDOQI (2006) Clinical practice guidelines and clinical practice recommendations for anemia in chronic kidney disease. Am J Kidney Dis 47[5 Suppl 3]:S11145 11. K/DOQI (2002) Clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis 39(2 Suppl 1):S1S266 12. Fadrowski JJ, Pierce CB, Cole SR, Moxey-Mims M, Warady BA, Furth SL (2008) Hemoglobin decline in children with chronic kidney disease: baseline results from the chronic kidney disease in children prospective cohort study. Clin J Am Soc Nephrol 3 (2):457462 13. United States Growth Charts. Available at: http://www.cdc.gov/ nchs/about/major/nhanes/growthcharts/datafiles.htm 14. Koshy SM, Geary DF (2008) Anemia in children with chronic kidney disease. Pediatr Nephrol 23(2):209219 15. Bamgbola OF, Kaskel FJ, Coco M (2009) Analyses of age, gender and other risk factors of erythropoietin resistance in pediatric and adult dialysis cohorts. Pediatr Nephrol 24(3):571579 16. Port RE, Mehls O (2009) Erythropoietin dosing in children with chronic kidney disease: based on body size or on hemoglobin deficit? Pediatr Nephrol 24(3):435437 17. Rao DS, Shih MS, Mohini R (1993) Effect of serum parathyroid hormone and bone marrow fibrosis on the response to erythropoietin in uremia. N Engl J Med 328(3):171175 18. Kruse A, Uehlinger DE, Gotch F, Kotanko P, Levin NW (2008) Red blood cell lifespan, erythropoiesis and hemoglobin control. Contrib Nephrol 161:247254 19. Nangaku M, Eckardt KU (2006) Pathogenesis of renal anemia. Semin Nephrol 26(4):261268 20. KDOQI (2007) Clinical Practice Guideline and Clinical Practice Recommendations for anemia in chronic kidney disease: 2007 update of hemoglobin target. Am J Kidney Dis 50(3):471530 21. McClellan W, Aronoff SL, Bolton WK, Hood S, Lorber DL, Tang KL, Tse TF, Wasserman B, Leiserowitz M (2004) The prevalence of anemia in patients with chronic kidney disease. Curr Med Res Opin 20(9):15011510 22. McFarlane SI, Chen SC, Whaley-Connell AT, Sowers JR, Vassalotti JA, Salifu MO, Lis S, Wang C, Bakris G, McCullough PA, Collins AJ, Norris KC (2008) Prevalence and associations of anemia of CKD: Kidney Early Evaluation Program (KEEP) and National Health and Nutrition Examination Survey (NHANES) 1999-2004. Am J Kidney Dis 51[4 Suppl 2]:S46S55 23. Cruzado JM, Rico J, Grinyo JM (2008) The renin angiotensin system blockade in kidney transplantation: pros and cons. Transpl Int 21(4):304313

prevalence of both corrected and uncorrected anemia in children with CKD as disease stage advances. Among children with stage 3 CKD, even after controlling for estimated GFR, factors associated with a significantly increased risk of anemia at 12 month follow-up include age >2 years, male sex, earlier era of registration into the NAPRTCS cohort, and treatment with anti-hypertensive medications at baseline. However, the presence of anemia, whether successfully treated or not, was not associated with an increased risk of progression to ESRD among children of this cohort with stage 3 CKD.
Acknowledgment Dr. Atkinson was supported by grant funding from Amgen, The Thrasher Research Fund, The National Kidney Foundation of Maryland, and The National Kidney Foundation.

References
1. Mitsnefes MM, Kimball TR, Kartal J, Witt SA, Glascock BJ, Khoury PR, Daniels SR (2006) Progression of left ventricular hypertrophy in children with early chronic kidney disease: 2-year follow-up study. J Pediatr 149(5):671675 2. Schaefer F (2008) Cardiac disease in children with mild-tomoderate chronic kidney disease. Curr Opin Nephrol Hypertens 17(3):292297 3. Gerson A, Hwang W, Fiorenza J, Barth K, Kaskel F, Weiss L, Zelikovsky N, Fivush B, Furth S (2004) Anemia and healthrelated quality of life in adolescents with chronic kidney disease. Am J Kidney Dis 44(6):10171023 4. Furth SL, Cole SR, Fadrowski JJ, Gerson A, Pierce CB, Chandra M, Weiss R, Kaskel F (2007) The association of anemia and hypoalbuminemia with accelerated decline in GFR among adolescents with chronic kidney disease. Pediatr Nephrol 22(2):265271 5. Wuhl E, Schaefer F (2008) Therapeutic strategies to slow chronic kidney disease progression. Pediatr Nephrol 23(5):705716 6. Filler G, Mylrea K, Feber J, Wong H (2007) How to define anemia in children with chronic kidney disease? Pediatr Nephrol 22(5):702707 7. Staples AO, Wong CS, Smith JM, Gipson DS, Filler G, Warady BA, Martz K, Greenbaum LA (2009) Anemia and risk of hospitalization in pediatric chronic kidney disease. Clin J Am Soc Nephrol 4(1):4856 8. Amaral S, Hwang W, Fivush B, Neu A, Frankenfield D, Furth S (2006) Association of mortality and hospitalization with achievement of adult hemoglobin targets in adolescents maintained on hemodialysis. J Am Soc Nephrol 17(10):28782885