Abstract
Background Poor linear growth is a consequence of chronic kidney disease (CKD) that has been linked to adverse outcomes. Metabolic acidosis (MA) has been identified as a risk factor for growth failure. We investigated the longitudinal relationship between MA and linear growth in children with CKD and examined whether treatment of MA modified linear growth.
Methods To describe longitudinal associations between MA and linear growth, we used serum bicarbonate levels, height measurements, and standard deviation (z scores) of children enrolled in the prospective cohort study Chronic Kidney Disease in Children. Analyses were adjusted for covariates recognized as correlating with poor growth, including demographic characteristics, glomerular filtration rate (GFR), proteinuria, calcium, phosphate, parathyroid hormone, and CKD duration. CKD diagnoses were analyzed by disease categories, nonglomerular or glomerular.
Results The study population included 1082 children with CKD: 808 with nonglomerular etiologies and 274 with glomerular etiologies. Baseline serum bicarbonate levels ≤22 mEq/L were associated with worse height z scores in all children. Longitudinally, serum bicarbonate levels ≤18 and 19–22 mEq/L were associated with worse height z scores in children with nonglomerular CKD causes, with adjusted mean values of –0.39 (95% CI, –0.58 to –0.2) and –0.17 (95% CI, –0.28 to –0.05), respectively. Children with nonglomerular disease and more severe GFR impairment had a higher risk for worse height z score. A significant association was not found in children with glomerular diseases. We also investigated the potential effect of treatment of MA on height in children with a history of alkali therapy use, finding that only persistent users had a significant positive association between their height z score and higher serum bicarbonate levels.
Conclusions We observed a longitudinal association between MA and lower height z score. Additionally, persistent alkali therapy use was associated with better height z scores. Future clinical trials of alkali therapy need to evaluate this relationship prospectively.
Introduction
Linear growth impairment is a consequence of CKD that has been associated with profound risk for adverse outcomes (1⇓–3). In an early investigation from the Pediatric Growth and Development Special Study, every 1 SD decrease in height was an associated 14% increased risk for death (1). Similarly, in the North American Pediatric Renal Trials and Collaborative Studies, compared with children with heights at or above the 1st percentile, children with heights below the 1st percentile had a two-fold higher risk of death (4). Poor growth also has profound psychosocial effects. Children with CKD who suffer from short stature report lower physical functioning scores on health-related quality of life assessment tools (2,3). Higher parental scores of physical and social functioning have been associated with increases in height z score (3). Given that growth failure is estimated to affect up to 35% of the pediatric CKD population (5), it is important to understand factors that contribute to short stature in order to manage these patients better.
The etiology of growth failure in pediatric CKD is complex, but numerous studies point to metabolic acidosis (MA) as a contributing factor (6⇓⇓⇓⇓–11). The theorized mechanism involves disturbances of growth hormone (GH), and its mediating hormone, IGF-1 (12⇓–14). MA has been reported to impair GH secretion, reduce hepatic IGF-I mRNA, and alter concentration of and sensitivity to IGF-1. Baseline data from the Chronic Kidney Disease in Children (CKiD) study show that as many as one third of children with mild to moderate CKD have low serum bicarbonate levels, a proxy for MA (7,15,16). In a cross-sectional analysis of the CKiD cohort, Rodig et al. found that children with a serum bicarbonate of <18 mEq/L had a height SD score that was 0.67 lower than for those with a serum bicarbonate ≥22 mEq/L (95% confidence interval [95% CI], –1.03 to –0.31) (7). Additional support for the negative effect of MA on linear growth in pediatric CKD is the extremely high prevalence of growth failure in children with renal tubular acidosis (RTA) (9⇓–11). However, it must be stated that these studies were small in number and largely only included children with normal eGFR. Alkali therapies (bicarbonate, citrate-containing solutions), used for the treatment of chronic MA, are widely available and well tolerated (17). Despite the reported association between MA and linear growth impairment and the availability of tolerable treatment options, treatment rates of MA remain suboptimal, with only one third of children with low serum bicarbonates in CKiD reporting treatment with alkalinizing agents (7,16). Given the prevalence of MA, low reported rates of treatment, and the profound effect of growth failure, longitudinal studies that include children with impaired kidney function are needed to inform management practices better. In this study, we described and characterized the longitudinal relationship between low serum bicarbonate, a surrogate for MA, and linear growth in children enrolled in the CKiD study. We hypothesized that low serum bicarbonate would be associated with lower height z scores and that treatment of low bicarbonate with alkali therapy would be associated with improved height.
Materials and Methods
Study Population
The CKiD study is a longitudinal observational cohort study aimed at investigating and characterizing the effect of CKD progression in children. There are more than 50 enrollment institutions across North America. Study eligibility requires that children carry a diagnosis of CKD and mild to moderately impaired kidney function defined as an eGFR of 30–90 ml/min per 1.73 m2. CKiD has enrolled children from 6 months to 16 years old at study entry; our study required a minimum age of 2 years old (minimum age for standing height measurements). In the first year, participants are seen twice and then annually thereafter. At each study visit, demographic and clinical data are obtained, including growth measurements and serum samples for measurement of kidney function and related biomarkers. A full description of the CKiD study and cohort has been previously published (18). All participants and families provided informed assent or consent. All protocols were approved by the Institutional Review Board.
Primary Exposure: Serum Bicarbonate
Serum bicarbonate results were obtained and measured at local study site laboratories. Low serum bicarbonate was defined as ≤22 mEq/L, and normal was defined as >22 mEq/L (19). For both baseline and longitudinal analyses, abnormal serum bicarbonate was further clinically categorized as ≤18 mEq/L (very low) and 19–22 mEq/L (low). We also looked at height z score on serum bicarbonate as a continuous predictor in specified analyses.
Primary Outcome: Linear Growth
A wall-mounted stadiometer was used to measure height (i.e., linear growth) at study visits. Final recorded height was based on averaging two separate measurements to the nearest 0.1 cm. If the measures differed by >0.3 cm, a third measurement was made, and an average of all three measurements used. Height was converted to height z scores (i.e., standard deviations) and percentiles according to Centers for Disease Control and Prevention estimates for the normal population adjusted for age and sex (20). Longitudinal analyses included participant visits during regular study follow-up among those <20 years of age with complete data on serum bicarbonate (exposure) and height z score (primary outcome).
Stratification and Covariate Definitions
The CKD diagnoses were broadly classified into two primary disease categories: nonglomerular or glomerular (specific CKD diagnoses are described in Supplemental Table 1), and all analyses were stratified as such. In longitudinal analyses, participants were also stratified by GFR ≥45 ml/min per 1.73 m2 (mild to moderate CKD) and <45 ml/min per 1.73 m2 (moderate to severe CKD).
Analyses were adjusted for covariates known to affect kidney disease progression—a variable recognized as correlating with poor linear growth (7,15,21). Covariates included demographic variables such as age, sex, abnormal birth history (defined as premature birth, low birth weight, or small for gestational age), and mid-parental height (defined by biologic parents’ height measured at baseline study visit). Clinical covariates included GFR (calculated using the CKiD U25 eGFR equation and iohexol measured, where available) (22), CKD duration (in years), proteinuria (defined categorically on the basis of urine/protein creatinine ratio, <0.5, 0.5–<2, or ≥2 mg/mg), intact parathyroid hormone (treated as continuous), and serum calcium and phosphate levels (classified as abnormal on the basis of age-specific thresholds). In longitudinal analyses evaluating the effect of a history of alkali therapy treatment, eGFR and proteinuria were treated as a “lagged” variable such that values from the previous study visit were used. “Lagged” values were used in order to provide clinical evidence as to how historical variables could predict the subsequent year’s outcomes.
Missing covariate data were imputed using multiple imputation by chained equations methods to limit the effect of missing follow-up data. The method used Gibbs sampling to perform five imputations of missing values for the “target covariate” on the basis of values from all other covariates in the dataset. Missing values were imputed separately for those with nonglomerular and glomerular diagnoses.
Statistical Analyses
Medians, interquartile ranges, and proportions described the demographic, clinical history, growth, and kidney disease characteristics of the cohort at participants’ first available visit.
To characterize the association between serum bicarbonate and linear growth, we used repeated measures linear regression models with height z score as the outcome and bicarbonate from the previous year as a categorical exposure. Models were stratified by diagnosis and were unadjusted (i.e., no covariates), partially adjusted (specifically, age, sex, abnormal birth history, mid-parental height, and previous levels of eGFR and proteinuria), and fully adjusted (the same covariates with the addition of calcium, phosphate, intact parathyroid hormone, and CKD duration). Generalized estimating equations (GEE) were used to account for longitudinal measurements within an individual. As a supplementary analysis, we investigated serum bicarbonate as a predictor of growth velocity z scores.
Additionally, we characterized the relationship between serum bicarbonate and height z score among participants aged ≤13 years (i.e., prepubertal age range) by longitudinal alkali therapy use. Specifically, the unit of analysis was pairs of visits. We restricted to participants who reported using alkali therapy at the first visit, and we compared those who discontinued use (i.e., used alkali therapy at the previous visit but not at the current visit) with those who were persistent users (i.e., used alkali therapy at both the previous and current visit). Children with varying alkali therapy use during yearly follow-up (i.e., discontinued user during follow-up who became a persistent user or persistent user who became a discontinued user) could contribute data to both groups. GEE were also used to account for correlated repeated measures within an individual. Estimates of differences are presented with two-sided 95% CIs. Differences were statistically significant if the interval did not contain the null value (0), which corresponds to P<0.05. All analyses were conducted using R v4.0.2 (The R Foundation for Statistical Computing, Vienna, Austria). To address missing data, multiple imputation on the basis of predictive mean matching for all covariates was used with the mice function of the “mice” package (v3.13.0). Linear models with GEE were calculated using the geeglm function of the “geepack” package (v1.3–1).
Results
The study population comprised 1082 children: 808 with nonglomerular etiologies of their CKD and 274 with glomerular causes of their CKD. Multivariate analyses were performed to examine the association between baseline demographic, biochemical, and clinical variables with serum bicarbonate values. Descriptive characteristics of the study participants are detailed in Tables 1 and 2. In both primary disease etiology categories, there was a predominance of patients who were White and boys, and low bicarbonate was associated with lower eGFR and more significant proteinuria. Abnormal birth history was similar among all bicarbonate groups, independent of CKD etiology. In both primary disease groups, baseline serum bicarbonate ≤22 mEq/L was associated with lower height measurements, worse height z scores, and more patients on GH therapy. Regarding alkali therapy, in patients with nonglomerular disease, serum bicarbonate ≤22 mEq/L was associated with a borderline higher rate of reported alkali therapy use, whereas it trended in this direction in children with glomerular disease but was not statistically significant. For patients with nonglomerular diseases, 45% (360/808) had a baseline bicarbonate of ≤22 mEq/L, with 34% (122/360) of those patients reporting treatment with alkali therapy. For children with glomerular diseases, 35% (97/274) had a baseline bicarbonate of ≤22 mEq/L, and 16% (16/97) endorsed alkali therapy treatment. Reported GH therapy use was low in the overall analyzed cohort at 9%.
Descriptive characteristics of study population with a nonglomerular CKD diagnosis overall and by bicarbonate levels at baseline
Descriptive characteristics of study population with a glomerular CKD diagnosis overall and by bicarbonate levels at baseline
For longitudinal analyses, we grouped serum bicarbonate from the previous study visit (i.e., lagged values) by clinically relevant categories as very low (≤18 mEq/L), low (19–22 mEq/L), and normal (>22 mEq/L), and we examined the distribution of height z scores across bicarbonate groups. This was done to examine the clinical utility of bicarbonate levels in predicting future height outcomes. Figure 1 demonstrates that, longitudinally, worse serum bicarbonate levels were associated with worse height z scores in all children with CKD, independent of CKD etiology. In fully adjusted models, current serum bicarbonate and growth measurements were utilized. In these analyses, we continued to find that very low and low bicarbonates were associated with significantly worse height z scores in children with nonglomerular CKD (fully adjusted mean –0.39 [95% CI, –0.58 to –0.2] and –0.17 [95% CI, –0.28 to –0.05], respectively; Table 3). When restricted to patients who had measured GFR available, we found this association continued to persist and was more pronounced (Table 4). Height z scores were overall higher in children with glomerular CKD compared with those with nonglomerular CKD (Figure 1). In examining whether lower serum bicarbonate was linked to worse height z score in children with glomerular diseases, the only significant association noted was in unadjusted analyses of children with serum bicarbonate 19–22 mEq/L (Tables 3 and 4). In sensitivity analyses excluding the children who reported GH use, associations were unchanged for children with nonglomerular CKD. Due to the small number of children with a glomerular diagnosis, data from low bicarbonate groups were combined, and height z scores in children with serum bicarbonate ≤22 mEq/L were compared with children with normal bicarbonates, with no significant relationship observed (Supplemental Table 2).
Distribution of height z score by previous visit bicarbonate levels among person-visits contributed by participants. (A) Participants with a diagnosis of nonglomerular CKD. (B) Participants with a diagnosis of glomerular CKD.
Unadjusted, partially adjusted, and fully adjusted models of height z score on serum bicarbonate, using a categorical predictor
Unadjusted, partially adjusted, and fully adjusted models of height z score on serum bicarbonate, using a categorical predictor restricted to person-visits with directly measured GFR only
In analyses that were stratified by GFR, both estimated (Table 5) and measured (Table 6), using ≥45 ml/min per 1.73 m2 (mild to moderate CKD) and <45 ml/min per 1.73 m2 (moderate to severe CKD) categories, as potential effect modifiers of the relationship. In children with nonglomerular CKD causes, across all models using eGFR, low and very low serum bicarbonate were associated with significantly worse height z score in children with an eGFR <45 ml/min per 1.73 m2; the same association was seen with measured GFR in children with serum bicarbonate 19–22 mEq/L. This association reached significance across both measured GFR groups in children with serum bicarbonate ≤18 mEq/L. For children with glomerular diseases, in fully adjusted analyses, no association between serum bicarbonate and height z score was noted, independent of GFR type.
Unadjusted, partially adjusted, and fully adjusted models of height z score on serum bicarbonate, stratified by eGFR <45 and ≥45 ml/min per 1.73 m2, using a categorical predictor
Unadjusted, partially adjusted, and fully adjusted models of height z score on serum bicarbonate, stratified by measured GFR <45 and ≥45 ml/min per 1.73 m2, using a categorical predictor
We also evaluated the association between MA and linear growth, treating serum bicarbonate as a continuous variable. Supplemental Table 3 depicts differences in height z score per 1 mEq/L higher serum bicarbonate in the previous year. In fully adjusted models, we observed that a 1 mEq increase in serum bicarbonate was associated with increases in height z scores in all children with CKD; however, this relationship was not significant.
In a categorical exploratory analysis of the entire cohort restricted to prepubertal aged children, low serum bicarbonate levels were linked to lower height z scores, achieving significance in the very low serum bicarbonate category when eGFR was used for fully adjusted models. When measured GFR was utilized, this relationship was significant in fully adjusted models for both very low and low serum bicarbonates (Supplemental Tables 4 and 5).
Finally, we investigated the effect of treatment of MA on height z score in children of prepubertal age (Tables 7 and 8). We examined current height z score as a function of change in serum bicarbonate in participants reporting alkali therapy use at the previous study visit. Height outcome was separated by current reported alkali therapy use (i.e., “persistent” use if still being treated with alkali therapy versus “discontinued” use for those no longer reporting use of alkali therapy). In these models that utilized bicarbonate in a “lagged” manner, persistent alkali therapy users had a significant positive association between their height z score and serum bicarbonate levels; the significance of this association was lost when measured GFR was utilized. Independent of the GFR used, there was a positive but nonsignificant relationship in discontinued users.
Unadjusted and adjusted models of height z score on previous visits serum bicarbonate stratified by discontinued and persistent alkali therapy use on the basis of two annual study visits
Unadjusted and adjusted models of height z score on previous visits serum bicarbonate stratified by discontinued and persistent alkali therapy use on the basis of two annual study visits
Discussion
Using annual serum bicarbonate values over a robust duration of follow-up, our data suggest a longitudinal association between MA and lower height z score. After adjusting for demographic characteristics, markers of CKD severity, and pertinent clinical variables, serum bicarbonate ≤22 mEq/L was associated with lower height z scores, with the worst height z scores observed in the lowest bicarbonate category (≤18 mEq/L). This association reached significance among children with nonglomerular CKD only. Although overall height z scores were reduced in all children with CKD, children with nonglomerular diseases had greater deficits in height z score than children with glomerular diseases. Not unexpectedly, use of measured GFR data showed similar associations compared with use of eGFR except when measured GFR was dichotomized as >45 and ≤45 ml/min per 1.73 m2. In these analyses, in participants with nonglomerular diseases, very low serum bicarbonate was associated with worse height z score across both measured GFR groups compared with analyses that utilized eGFR where this association was only seen in children in the ≤45 ml/min per 1.73 m2 group. This is noteworthy because although use of measured GFR is not routine clinical practice, it is more accurate than eGFR, suggesting that the association of worse height z scores with MA may be present in the milder CKD group. Finally, and of clinical relevance, our data suggest that alkali therapy use as a marker of treatment of MA was associated with improved height z score, particularly in persistent users.
We found height outcome differences between children with nonglomerular and glomerular etiologies of CKD. Observed differences in linear growth could be attributed to sample size differences between the groups, later age of CKD onset in children with glomerular disease (i.e., less time during which the sequelae of CKD can affect active linear growth), older age of the participants with glomerular disease, and previously published evidence that indicates these primary disease groups may not be similarly affected by CKD comorbidities (16,23).
Our findings are, in part, in line with previous baseline investigations of the relationship between low serum bicarbonate and growth. In a prior cross-sectional study using the CKiD cohort, Rodig et al. observed that baseline height was lower in children with a baseline serum bicarbonate of <18 mEq/L compared with those whose serum bicarbonate was ≥22 mEq/L (7). They did not find a link between serum bicarbonate ≥18 to <22 mEq/L and worse height outcomes, unlike in our study where serum bicarbonate 19–22 mEq/L was associated with greater deficits in height in children with nonglomerular CKD. Potential reasons to account for this difference may be due to the longitudinal nature of our analyses, the stratification of our data by CKD etiology (although their study did adjust for CKD diagnosis), and differences in bicarbonate categorization values. Although noting the potential benefits of alkali therapy on height outcomes, unlike in our study, differences in height on the basis of alkali therapy use were not tested. Our study is the first to incorporate robust longitudinal data in a multi-ethnic cohort to investigate this association. Our novel evaluation of height on the basis of alkali therapy uses points to the potentially positive effect of treatment of MA on height.
Although longitudinal evaluations of the relationship between MA and growth are small in number, data from a European, prospective, observational study of children post kidney transplant found that MA severity predicted poor linear height in pediatric kidney transplant recipients (24). This relationship has also been investigated in the European cohort of children with moderate to severe CKD enrolled in the Comorbidity in Children with CKD (4C) study (8). Using 4C data, Harambat et al. found a trend-level association between serum bicarbonate <18 mmol/L and height SD score in univariate analyses. In multivariate analyses, the negative association persisted, although significance was lost. It is possible that inclusion of children with more severe CKD at enrollment (compared with mild to moderate CKD required for CKiD study entry) may have contributed to observed differences between this study and our investigation. CKD severity has been shown to correlate negatively with height outcomes (7,15,21). Thus, it is possible that the potential contribution of MA to poor linear growth may not have been as clearly defined in the presence of stronger correlates of poor growth such as kidney function impairment. Another important consideration is the study population characteristics. In the United States, CKD is up to three times more prevalent in African American children. So, it may be difficult to draw accurate inferences from European cohorts and apply them to a diverse North American population (4). Our investigation incorporates longitudinal data using a diverse population of children, which may better represent associations between MA and linear growth in pediatric CKD and the efficacy of alkali therapy to improve growth.
The growth benefits of alkali therapy have mostly been shown in older, small studies of children with RTA and normal eGFR and, therefore, mild CKD (9⇓–11,25). Alkali therapies are widely available and well tolerated in the pediatric CKD population (17). Despite the availability of tolerable treatment options, published benefits of acidosis correction in growth impaired children with RTA, and proposed benefits on overall CKD progression (16,26⇓⇓⇓⇓–31), treatment rates of MA remain suboptimal, with only approximately one third of children with low serum bicarbonates in CKiD (one of the largest cohorts of children with CKD) reporting treatment with alkalinizing agents (7,16). Data from our current investigation suggest that long-term use of alkali therapy may have beneficial effects on height in children with CKD. However, clinical trials of alkali therapy in children with varying severity of impaired kidney function are needed to inform practitioners better of the potential benefits of treatment, given that no such trial exists to date.
Although our study has several strengths, there are limitations. Nutritional data in this cohort were incomplete. So, we were unable to account fully for its effect on height. We do include data on underweight children (on the basis of body mass index; Tables 1 and 2). Only 4% of the entire cohort is underweight, reflecting the likely low occurrence of severe malnutrition. Serum bicarbonate was used as an indicator of MA in this study because serum pH data were not available; it is possible that serum bicarbonate was not equivalent to actual acid/base status. Additionally, we were not able to assess true duration of alkali therapy because exact start dates were unknown, and for those with historical use, we were unable to determine when the alkali agent was discontinued. Important to note is that we were unable to confirm adherence with alkali therapy in those who reported use. Another limitation is that we were also only able to account for the previous year’s level of bicarbonate and other covariates measured one year prior. It would have been preferable to account for longer duration of clinical covariates in a marginal structural model framework, but data were limited to account for longer than one year earlier. An additional limitation is the inclusion of the small number of children on GH in primary analyses; however, sensitivity analyses showed no significant changes in study conclusions when these patients were excluded. Finally, we are aware that results may be affected by confounding by indication because children with more severe acidosis and complications were more likely to be prescribed alkali therapy.
Despite its limitations, to our knowledge, our study is the first to examine the longitudinal relationship between MA and linear growth, and the potential effect of acidosis correction, in a multi-ethnic cohort of children with varying CKD severity. Although there are safe and effective therapies to treat MA, an increased understanding of this relationship may inform treatment practices and prove crucial to improving pediatric CKD care. Although our observed associations were small in number, our findings of a negative correlation between low serum bicarbonate and linear growth in children with CKD, and the suggested height benefits of alkali therapy are important, given the profound effect impaired growth may have in this vulnerable population. Future clinical trials of alkali therapy need to evaluate this relationship and other important disease outcomes prospectively in children with MA and chronically impaired kidney function.
Disclosures
M. Carroll reports ownership interest in Verily Life Sciences and research funding from Verily Life Sciences. A. Dauber reports ownership interest in Ascendis, Biomarin, and Novo Nordisk and research funding from Biomarin. L.A. Greenbaum reports consultancy for Abbvie, Advicenne, Alexion, Arrowhead Pharmaceuticals, Aurinia, CorMedix, NephroDI Therapeutics, Natera, Novartis, Roche, and Otsuka; research funding from Abbvie, Advicenne, Alexion, Apellis, Aurinia, Horizon Pharmaceuticals, Reata Pharmaceuticals, and Vertex; honoraria from Alexion; an advisory or leadership role with Alexion; and other interests or relationships (DSMB payments) with Akebia, Alnylam, Relypsa, Travere, and UCSD. F.J. Kaskel reports research funding from NIDDK and an advisory or leadership role for Frost Valley YMCA and Nephcure, Inc. M.L. Melamed reports an advisory or leadership role for the American Board of Internal Medicine Nephrology Exam Committee and has other interests in or relationships with the American Society of Nephrology and the New York Society of Nephrology. D.K. Ng reports consultancy for Ashvattha Therapeutics. B.A. Warady reports consultancy for Amgen, Bayer, Lightline Medical, Reata, Relypsa, and UpToDate; research funding from Baxter Healthcare; honoraria from Amgen, Bayer, Reata, Relypsa, and UpToDate; and an advisory or leadership role with the Midwest Transplant Network Governing Board, the National Kidney Foundation, North American Pediatric Renal Trials and Collaborative Studies, and the NTDS Board of Directors. All remaining authors have nothing to disclose.
Funding
This work was supported by the Developmental and Translational Nephrology Training Grant (T32 DK007110) and the Eunice Kennedy Shriver National Institute of Child Health and Human Development (K12HD001399).
Acknowledgments
Data in this manuscript were collected by the Chronic Kidney Disease in Children prospective cohort study (CKiD) with clinical coordinating centers (Principal Investigators) at Children’s Mercy Hospital and the University of Missouri–Kansas City (Bradley Warady) and Children’s Hospital of Philadelphia (Susan Furth), Central Biochemistry Laboratory (George Schwartz) at the University of Rochester Medical Center, and data coordinating center (Alvaro Muñoz and Derek Ng) at the Johns Hopkins Bloomberg School of Public Health. The CKiD Study is funded by the National Institute of Diabetes and Digestive and Kidney Diseases, with additional funding from the National Institute of Child Health and Human Development, and the National Heart, Lung, and Blood Institute (U01-DK-66143, U01-DK-66174, U24-DK-082194, U24-DK-66116). The authors thank Ankur Patel for assistance in preparing revisions of this manuscript. The CKID website is located at https://www.statepi.jhsph.edu/ckid.
Author Contributions
D.D. Brown, F.J. Kaskel, and M.L. Melmed weres responsible for funding acquisition; All authors wrote the original draft of the manuscript; D.D. Brown, M. Carroll, A. Dauber, M.L. Melamed, and D.K. Ng curated the data; D.D. Brown and A. Dauber were responsible for the investigation; D.D. Brown, A. Dauber, and M.L. Melamed were responsible for supervision; D.D. Brown, S.L. Furth, M.L. Melamed, and B.A. Warady conceptualized the study; M. Carroll and D.K. Ng were responsible for validation; M. Carroll and D.K. Ng were responsible for the formal analysis and software; B.A. Warady and S.L Furth were responsible for resources; and all authors were responsible for the methodology and reviewed and edited the manuscript.
Supplemental Material
This article contains supplemental material online at http://kidney360.asnjournals.org/lookup/suppl/doi:10.34067/KID.0005402021/-/DCSupplemental.
Supplemental Table 1. Distribution of diagnoses within nonglomerular and glomerular participants.
Supplemental Table 2. Unadjusted, partially adjusted, and fully adjusted models of height z score on serum bicarbonate, using a categorical predictor among non-rGH users. Missing data were imputed for covariates in the partially and fully adjusted models.
Supplemental Table 3. Unadjusted, minimally adjusted, and fully adjusted models of height z score on previous visits’ serum bicarbonate. Generalized estimating equations used to account for repeated measures within an individual.
Supplemental Table 4. Among children <13 with adjustment of glomerular/nonglomerular diagnosis as a covariate and eGFR. Missing data were imputed for covariates in the partially and fully adjusted models.
Supplemental Table 5. Among children <13 with adjustment of glomerular/nonglomerular diagnosis as a covariate and measured GFR. Missing data were imputed for covariates in the partially and fully adjusted models.
Supplemental Summary 1. List of principal site investigators of the Chronic Kidney Disease in Children (CKiD) cohort study.
Footnotes
See related editorial, “Impact of Acidosis on Growth in Pediatric Chronic Kidney Disease: What Is the Current Evidence?,” on pages 590–596.
- Received August 17, 2021.
- Accepted January 10, 2022.
- Copyright © 2022 by the American Society of Nephrology
References
Jump to section
More in this TOC Section
Original Investigation
Cited By...
- No citing articles found.