Abstract
Background and objectives Metabolic abnormalities and cardiovascular disease (CVD) risk factors have rarely been systematically assessed in children with chronic kidney disease (CKD). We examined the prevalence of various CKD sequelae across the GFR spectrum.
Design, setting, participants, & measurements Data were used from 586 children participating in the Chronic Kidney Disease in Children (CKiD) study (United States and Canada) with GFR measured by iohexol plasma disappearance. Laboratory values and CVD risk factors were compared across GFR categories and with an age-, gender-, and race-matched community sample.
Results CKiD participants were 62% male, 66% Caucasian, 23% African American, and 15% Hispanic with a median age of 11 years and a median GFR of 44 ml/min per 1.73 m2. Compared with those with a GFR ≥ 50 ml/min per 1.73 m2, having a GFR < 30 ml/min per 1.73 m2 was associated with a three-fold higher risk of acidosis and growth failure and a four- to five-fold higher risk of anemia and elevated potassium and phosphate. Median GFR change was −4.3 ml/min per 1.73 m2 and −1.5 ml/min per 1.73 m2 per year in children with glomerular and nonglomerular diagnoses, respectively. Despite medication and access to nephrology care, uncontrolled systolic hypertension was present in 14%, and 16% had left ventricular hypertrophy. Children with CKD frequently were also shorter and had lower birth weight, on average, compared with norms.
Conclusions Growth failure, metabolic abnormalities, and CVD risk factors are present at GFR >50 ml/min per 1.73 m2 in children with CKD and, despite therapy, increase in prevalence two- to four-fold with decreasing GFR.
Introduction
Chronic kidney disease (CKD) affects an estimated 9% to 13% of the general U.S. population (1). More than $23.9 billion was spent on ESRD in 2007 (2). Understanding the occurrence of comorbid conditions as GFR declines is crucial to improving secondary prevention efforts and reducing ESRD costs. Although children compose a small proportion of the population with CKD, they suffer from many of the same comorbidities as adults. Studying children with primary kidney disease should yield insights into the relationship between kidney function decline and the development of metabolic abnormalities and cardiovascular risk. Almost 35% of mortality in young adults with CKD is attributed to cardiovascular disease (2). Few systematic data exist on CKD in children and youth to determine the threshold of kidney function impairment at which metabolic problems develop. In this report, we present clinical and demographic characteristics of the Chronic Kidney Disease in Children (CKiD) study by baseline category of measured GFR and compare them with population-based data using the National Health and Nutrition Examination Survey (NHANES) III.
Study Population and Methods
Details of the CKiD study design have been previously published (3). Briefly, eligible children aged 1 to 16 years with a Schwartz-estimated GFR (4) between 30 and 90 ml/min per 1.73 m2 were enrolled at 48 participating tertiary care pediatric nephrology programs across the United States and two sites in Canada. The baseline study visits that form the basis of this report occurred between January 2005 and August 2009, with the 1-year follow-up visits occurring between January 2006 and October 2010. Institutional Review Boards for each participating site approved the study protocol.
Family history of kidney disease and cardiovascular disease (CVD), past medical history, type of health insurance, birth history information, medication use, symptoms, and underlying CKD diagnosis and duration were obtained by parent report. Height and weight were determined by averaging three repeated measures via stadiometer and standing scale and percentiles, body mass index (BMI), height velocity, and growth failure (defined as height <3rd percentile for age and gender) were calculated. Pubertal (Tanner) stage was obtained by physician assessment. Clinical BP measurement technique using a handheld aneroid sphygmomanometer (Mabis MedicKit 5, Mabis Healthcare, Waukegan, IL) was standardized by centralized training with certified staff. Clinical BP was taken as the average of three readings and classified according to the National High Blood Pressure Education Program Fourth Report (5). Hypertension (HTN) was defined as BP (systolic or diastolic) ≥95th percentile for age, gender, and height (uncontrolled) or as self-report of HTN plus current treatment with antihypertensive medications. At the baseline and first annual follow-up visits, GFR was determined by directly measured plasma iohexol (GE Healthcare, Amersham Division, Princeton, NJ) disappearance curves; details of the GFR assessment method have been published previously (6). An estimated GFR value was used when a directly measured value was unavailable (7). Basic metabolic profile, including measurement of creatinine, was assessed using an enzymatic method on the Bayer Advia 2400 analyzer (Siemens Diagnostics, Tarrytown, NY). Inflammation was assessed with wide-range C-reactive protein (wrCRP). Complete blood count was measured locally. Cystatin C was measured using a turbidometric assay (Cystatin C Kit K0071, Dako Denmark A/S, Glostrup, Denmark).
At the first annual follow-up visit, serum total cholesterol (TC), triglycerides (TG), and HDL-cholesterol (HDL-C) were measured using the Bayer Advia 2400 analyzer after an overnight fast. Non-HDL cholesterol was calculated as the difference between TC and HDL-C. Ambulatory BP (ABP) was measured using the SpaceLabs 90217 device. Normative data for ABP and the definition of ambulatory HTN were based on American Heart Association recommendations on ABP in children and adolescents (8,9). Confirmed HTN is presence of clinical and ambulatory HTN, white coat HTN is clinical HTN in the presence of normal ABP (mean ABP <95th percentile and ABP load < 25%), and masked HTN is ambulatory HTN in the presence of normal clinical BP.
M-mode and Doppler echocardiograms were also obtained at the 1-year follow-up visit. Analyses of echocardiographic data were performed according to the American Society of Echocardiography criteria (10). Left ventricular mass was indexed (mass divided by height2.7 [g/m2.7]) to account for body size (11,12). Left ventricular hypertrophy (LVH) was defined as left ventricular mass index (LVMI) ≥ the 95th percentile.
To compare baseline family history and clinical and laboratory data of CKiD participants to population-based estimates, we used data from NHANES III, which used a complex, multistage, clustered sample of a civilian, noninstitutionalized population (13,14). A total of 32,106 children aged 1 to 17 years were examined. Of these, we analyzed 13,939 who had complete data on age, gender, and race. Not all laboratory measures were captured in the full range of ages for the children sampled in NHANES III.
Descriptive statistics (percentages and medians) were calculated for the CKiD cohort overall and by category of GFR (≥50, 40 to <50, 30 to <40, and <30 ml/min per 1.73 m2) at study entry with cutpoints determined based on the distribution of the CKiD sample. Trends across GFR categories were assessed using the Cochran–Armitage test and the nonparametric Cuzick test. Prevalence ratio estimates were obtained using Poisson regression with robust error variance (15,16).
To reflect the U.S. population, analyses of NHANES III data were weighted using the standard NHANES weights and adjusted for stratified and clustered sampling using SAS survey procedures. The estimates were also standardized to reflect the age, race, and gender distribution of the CKiD sample (17).
To evaluate GFR decline in the cohort, we examined longitudinal change in GFR from the baseline visit overall and by underlying glomerular or nonglomerular diagnosis. Longitudinal changes in GFR were estimated using linear regression on the individual GFR data in the log scale and expressed, after exponentiation, as annualized percent change.
All analyses were performed using SAS 9.2 (SAS Institute, Cary, NC); the accompanying figure was created using S-Plus 8.0 (Insightful, Seattle, WA).
Results
Characteristics of the 586 CKiD participants at study entry overall and by categories of measured GFR are listed in Table 1. Among the participants, there were 1055 directly measured and 609 estimated GFRs for determining annual percent decline. The median age was 11 years; the median GFR was 44 ml/min per 1.73 m2. Nineteen percent of the cohort had a history of low birth weight (<2500 g), 13% had a gestational age < 36 weeks, and 16% were small for gestational age. Among children enrolled at U.S sites, 49% had private insurance, 49% had public insurance, and only 2% were uninsured. Overall, 29% were treated with alkali therapy for acidosis, with the proportion increasing as GFR category declined (P < 0.01). Only 14% of the cohort were treated with erythropoiesis-stimulating agents (ESAs), and 29% were on iron supplementation. With the exception of ARB use, the proportions treated with these therapies increased with decreasing category of GFR (P < 0.05 for all comparisons).
Characteristics of 586 pediatric patients with CKD in the CKiD study
Compared with NHANES, children with CKD were substantially shorter but had a higher prevalence of being overweight. The median height and weight percentiles of the CKiD cohort overall were 24th and 46th, respectively; 24% of the CKiD cohort had a BMI > 90th percentile. Parental heights in CKiD were comparable to parental heights in NHANES (maternal height 163.3 cm in NHANES versus 162.6 cm in CKiD; paternal height 177.3 cm in NHANES versus 177.8 cm in CKiD). Within the CKiD cohort, child height and weight percentiles were cross-sectionally lower at lower GFRs. The percentage of individuals with BMI <15th percentile was significantly higher in the category with GFR <30 ml/min per 1.73 m2 (33%) as compared with the other GFR categories. The prevalence of growth failure increased with each decreasing category of GFR. Across categories of GFR, height velocity was constant.
Compared with children in the NHANES study, children in the CKiD study reported a higher percentage of either parent having HTN or stroke (29% versus 19%), high cholesterol (25% versus 21%), and diabetes (13% versus 4%). Parents of children in the CKiD study were slightly older; maternal and paternal ages were 37 and 40 years, respectively, in CKiD versus 35 and 39 years, respectively, in the NHANES study. BMI of fathers was higher in CKiD than NHANES (27.4 and 26.6), as was the BMI of CKiD mothers (27.1 versus 25.5 in NHANES).
Physical symptoms and complaints of decreased alertness (9%), loss of appetite (27%), weakness (27%), and leg pain (20%) were common and were generally highest in those with GFR <30 ml/min per 1.73 m2. Compared with children with GFR ≥50 ml/min per 1.73 m2 in the cohort, those with GFR 40 to <50, 30 to <40, or <30 ml/min per 1.73 m2 had increasing complaints of loss of appetite (21%, 27%, 31%, 32%, respectively; P = 0.02); weakness (21%, 23%, 34%, and 36%, respectively; P < 0.01), and leg pain (15%, 19%, 23%, and 25%, respectively; P = 0.02).
Table 2 shows the underlying causes of CKD comprising the 22% glomerular and 78% nonglomerular CKiD cohort. Children with underlying glomerular diagnoses were more likely to be African American (36% versus 20%), have a later age of onset of CKD (8 versus <1 year), were less likely to be low birth weight (13% versus 21%), were more likely to have systolic HTN (18% versus 13%), and had a higher urine protein-to-creatinine ratio (1.1 versus 0.4) compared with children with nonglomerular disease. Children with glomerular diagnoses were more likely to be entering puberty or postpubertal (56% Tanner III or greater versus 23% Tanner III or greater). For children aged 8 to 16 years, pubertal status was delayed for boys and girls compared with the NHANES cohort (Table 3), with more delay in boys than girls.
Causes of CKD in the CKiD cohort (n = 586)
Tanner staging in NHANES versus CKiD population (aged 8 to 16 years), CKiD n = 364
Median laboratory values for CKiD participants by GFR category and for the NHANES sample are presented in Table 4. As GFR category decreased, potassium and phosphate levels increased, as did urine protein-to-creatinine ratios, whereas bicarbonate (carbon dioxide) levels decreased despite increasing proportions of children on therapy, as demonstrated in Table 1. The median hemoglobin level in the CKiD subjects was 12.5 g/dl (interquartile range [IQR] 11.6 to 13.5), which was substantially lower than the normative values in NHANES, with 38% classified as anemic (defined by hemoglobin concentration <5th percentile of normal or current ESA use). The median wrCRP level was 0.12 mg/L in the CKiD cohort and did not increase with lower GFR categories; 58% had wrCRP levels ≤0.3 mg/L. Only 18% of the cohort exhibited wrCRP values >3 mg/L.
Laboratory characteristics of 586 pediatric patients with CKD
Median cholesterol, HDL-C, and non-HDL-C levels among children ≥4 years of age and TG among children ≥12 years of age in the CKiD cohort were 175, 48, 104, and 116 mg/dl compared with NHANES cohort medians of 161, 49, 89, and 78 mg/dl, respectively. Across categories of GFR within the CKiD cohort, non-HDL-C and TC were stable whereas TG levels increased markedly in the lower levels of GFR and HDL decreased significantly (Table 5).
Cardiovascular disease risk factors by level of GFR, n = 586
Despite 62% being on antihypertensive medication in the CKiD cohort, the median systolic and diastolic BP were at the 66th and 70th percentile for age, height, and gender, respectively. The values for BP were obtained at one visit. Most children (97%) had clinical BP measurements according to protocol at the baseline study visit; 20 children had two readings or fewer. Fourteen percent of CKiD participants had uncontrolled systolic HTN and 13% had uncontrolled diastolic HTN. The prevalence of systolic and diastolic HTN did not show a statistically significant increase across categories of GFR (Table 5). By ABP, 19% had masked and 13% had confirmed HTN. Thirty-five percent and 14% showed no evidence of normal nighttime systolic and diastolic dipping, respectively. Of 500 subjects who underwent baseline echocardiography, 16% had LVH. LVMI was similar by GFR category.
Table 6a shows the unadjusted prevalence ratios of several comorbid conditions by category of GFR using the GFR ≥50 ml/min per 1.73 m2 category as the reference. The prevalence of elevated potassium increased two times higher in the 40 to <50 ml/min per 1.73 m2 group and the 30 to <40 ml/min per 1.73 m2 group and four times higher in the ≤30 ml/min per 1.73 m2 group compared with the ≥50 ml/min per 1.73 m2 category. The risk of acidosis, elevated phosphate, growth failure, and anemia was two times higher. Table 6b shows the prevalence ratios by category of GFR excluding children on treatment for the particular condition.
Prevalence ratios by level of GFR
Prevalence ratios by level of GFR, excluding medication use
The overall annual percent decline was −4.2%, which corresponds approximately to a median decline of −1.8 ml/min per 1.73 m2 (IQR −6.6 to 1.6). The distribution of the annual percent decline in GFR is presented by nonglomerular and glomerular diagnosis in Figure 1. These percent declines correspond to approximate median absolute declines in GFR of −4.3 ml/min per 1.73 m2 (IQR −11.9 to 1.1) and −1.5 ml/min per 1.73 m2 (IQR −5.0 to 1.8) among those with underlying glomerular and nonglomerular diagnosis, respectively.
Annual percentage change in GFR across diagnosis categories.
Discussion
The CKiD study is the largest prospective cohort study of CKD in children in the United States and Canada. Compared with their healthy, population-based peers, children and adolescents in the CKiD cohort show height deficits across the entire range of GFR that become more pronounced with declining GFR levels. Compared with GFR ≥50 ml/min per 1.73 m2, prevalence of growth failure was two times higher with GFR 30 to <50 ml/min per 1.73 m2 and three times higher when GFR is <30 ml/min per 1.73 m2. In contrast, the distribution of weight percentiles was only shifted to lower values in the <30 ml/min per 1.73 m2 GFR group, suggesting that declining GFR affects caloric intake or metabolism. Similarly, the percentage of children with BMI <15th percentile increases substantially as GFR declines, as do symptoms of loss of appetite and weakness. In addition to well known risk factors for growth delay in children with CKD, including acidosis and abnormalities of the IGF–growth hormone axis, we found remarkably high rates of prematurity, being small for gestational age and having low birth weight, which may contribute to the height deficits and growth abnormalities seen in this population. Being small for gestational age has been previously reported as a risk factor for poor growth in healthy populations (18). Additionally, compared with normal children in the NHANES study, children aged 8 to 16 years also show evidence of markedly delayed puberty, particularly in boys.
Mild electrolyte abnormalities are present at relatively modest decrements in GFR. Potassium and phosphate levels increase with lower GFR. Despite an increasing proportion of children on alkali therapy, acidosis worsens with lower GFR. Metabolic acidosis contributes to poor growth in children, as well as loss of bone and muscle mass, and has been suggested to accelerate the progression of CKD (19). Hyperphosphatemia has been shown to be an independent risk factor for cardiovascular events and mortality in CKD (20). Our findings are consistent with data reported in the U.S. Renal Data System showing that patients with evidence of early kidney damage have undetected metabolic abnormalities (2).
Hemoglobin declines with decreasing GFR category (21), despite almost half of the children with GFR <30 ml/min per 1.73 m2 being treated with iron and/or ESAs. Low hemoglobin has been associated with increased mortality in children with ESRD (22), and targeting a higher hemoglobin in adults with CKD has been shown to improve LVMI (23). As shown in Table 4, levels of serum iron and transferrin saturation were normal and did not change across GFR categories, whereas ferritin levels increased somewhat with lower GFR, suggesting a pattern consistent with anemia of chronic disease rather than iron-deficiency anemia.
Compared with individuals with a GFR ≥50 ml/min per 1.73 m2, GFR <30 ml/min per 1.73 m2 was associated with a three-fold higher risk of acidosis and growth failure and a four- to five-fold higher risk of anemia and elevated potassium and phosphate. When we excluded children who were on treatment for each of these particular comorbid conditions, the point estimates for the prevalence ratios of each comorbid condition associated with lower category of GFR, including elevated potassium, acidosis, hyperphosphatemia, and anemia, became even larger, suggesting that our treatments are, to some extent, modifying the severity of these conditions in this cohort. Only in investigating growth failure did the exclusion of children treated with growth hormone actually make the association weaker. This may suggest that growth hormone is initiated only in those children with CKD who have the most extreme growth delay associated with CKD.
Prior studies in children with ESRD have shown that up to 31% of incident pediatric dialysis patients aged <19 years have experienced a cardiac-related event within 7 years of follow-up (24). Our data show that children with CKD have a remarkable constellation of risk factors for CVD even at a moderately decreased GFR of 40 to <50 ml/min per 1.73 m2. Compared with norms, they have a stronger family history of CVD, HTN, hypercholesterolemia, and diabetes. Despite having health insurance and access to nephrology specialty care, a significant proportion of children with CKD (13%) have uncontrolled systolic or diastolic HTN and evidence of target organ damage (LVH). A prior CKiD report showed an increase in LVH associated with masked and sustained HTN (25).
Elevation of TG and cholesterol levels was seen compared with normal children even in the group with GFR ≥ 50 ml/min per 1.73 m2. HDL was lower than norms even in the 40 to <50 ml/min per 1.73 m2 group, and non-HDL was also elevated. Prior reports in adults have demonstrated the independent risk of reduced estimated GFR and risk of death, cardiovascular events, and hospitalization in community-based populations. Go et al. showed a graded increased risk of cardiovascular events with decreasing estimated GFR below 60 ml/min per 1.73 m2, with a sharp increased risk associated with an estimated GFR <45 ml/min per 1.73 m2 (26). However, many of the adults studied had prior history of CVD and known risk factors for CVD and death. In our study of children and adolescents with primary kidney disease, it is likely that the CVD risk factors observed were all secondary to CKD. Our demonstration of dramatically higher numbers of CVD risk factors in a group of children with primary kidney disease suggests that impaired kidney function is a biologic risk factor for CVD. Although adult CKD and ESRD data had led us to expect an association between lower GFR and inflammatory markers, we did not find evidence of increased inflammation as measured by elevated wrCRP levels in our cohort. Our observation of increased family history of CVD among subjects with CKD suggests the possibility of common genetic risk factors for kidney disease and CVD.
Our study is the first to report rates of measured GFR decline in children and suggests a potential explanation for why a high proportion of young people present for medical care late in their course with stage 5 CKD with multiple CKD complications (2). Rapid decline in GFR and short duration with CKD may be associated with presentation at late-stage CKD or ESRD requiring dialysis (27). Our findings show more rapid decline in a subgroup with underlying glomerular disease, predominantly in African Americans. Higher incidence of presentation with late-stage CKD or ESRD among African Americans has long been noted.
In conclusion, children with moderate CKD show significant growth failure and pubertal delay that are prevalent even in individuals with a GFR >50 ml/min per 1.73 m2. Additionally, the constellation of CVD risk factors seen in these children with primary kidney disease suggests that decreased GFR is a causal factor in the excessive CVD mortality in individuals with kidney disease. In early-stage CKD, a clinical awareness of the heterogeneity of GFR decline according to CKD cause and more aggressive interventions to improve daytime and nighttime BP control, lowering phosphate intake, and treating acidosis and dyslipidemia may slow CKD progression and decrease the prevalence of CVD in children and young adults with CKD.
Disclosures
None.
Acknowledgments
The CKiD prospective cohort study is funded by the National Institute of Diabetes and Digestive and Kidney Diseases, with additional funding from the National Institute of Neurologic Disorders and Stroke; the National Institute of Child Health and Human Development; and the National Heart, Lung, and Blood Institute. The CKiD prospective cohort study has clinical coordinating centers (principal investigators) at Children's Mercy Hospital and the University of Missouri–Kansas City (Bradley Warady, MD) and Children's Hospital of Philadelphia and the University of Pennsylvania (Susan Furth, MD, PhD), a Central Biochemistry Laboratory at the University of Rochester (George Schwartz, MD), and a data coordinating center at the Johns Hopkins Bloomberg School of Public Health (Alvaro Muñoz, PhD) (U01-DK-66143, U01-DK-66174, U01-DK-082194, and U01-DK-66116).
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
- Received January 4, 2011.
- Accepted April 29, 2011.
- Copyright © 2011 by the American Society of Nephrology