Abstract
Importance
Most children with end-stage kidney disease (ESKD) are treated with dialysis prior to transplant. It is not known whether their outcomes have changed in recent years.
Objective
To determine if all-cause, cardiovascular, and infection-related mortality rates for children and adolescents beginning dialysis improved between 1990 and 2010.
Design, Setting, and Participants
Retrospective cohort study of patients younger than 21 years initially treated with dialysis for ESKD, recorded in the United States Renal Data System between 1990 and 2010. Children with a prior kidney transplant were excluded. We used Cox proportional hazard models to estimate the hazard ratios (HRs) for mortality associated with a 5-year increment in year of ESKD treatment initiation. Primary analyses censored observation at kidney transplant.
Main Outcomes and Measures
All-cause, cardiovascular, and infection-related mortality.
Results
A total of 23 401 children and adolescents who initiated ESKD treatment with dialysis at younger than 21 years between 1990 and 2010 were identified. Crude mortality rates during dialysis treatment were higher among children younger than 5 years at the start of dialysis compared with those who were 5 years and older. Mortality rates for both children and adolescents being treated for ESKD with dialysis decreased significantly between 1990 and 2010.
| 1990–1994 | 1995–1999 | 2000–2004 | 2005–2010 | Overall (1990–2010) | |
|---|---|---|---|---|---|
| Age <5 y at ESKD Treatment Initiation | |||||
| Total No. of patients | 692 | 734 | 845 | 1179 | 3450 |
|
| |||||
| Person-years, No. | 1613 | 1550 | 1670 | 2303 | 7136 |
|
| |||||
| All-cause deaths, No. | 181 | 151 | 181 | 192 | 705 |
|
| |||||
| Mortality per 1000 person-years | 112.2 | 97.4 | 108.4 | 83.4 | 98.8 |
|
| |||||
| Adjusted HR (95% CI) per 5-y increment in calendar year of ESKD initiation, 1990–2010 | 0.80 (0.75–0.85) | ||||
|
| |||||
| Age ≥5 y at ESKD Treatment Initiation | |||||
| Total No. of patients | 4368 | 4661 | 5103 | 5819 | 19 951 |
|
| |||||
| Person-years, No. | 14 595 | 15 749 | 15 350 | 13 104 | 58 799 |
|
| |||||
| All-cause deaths, No. | 651 | 661 | 618 | 340 | 2270 |
|
| |||||
| Mortality per 1000 person-years | 44.6 | 42.0 | 40.3 | 25.9 | 38.6 |
|
| |||||
| Adjusted HR (95% CI) per 5-y increment in calendar year of ESKD initiation, 1990–2010 | 0.88 (0.85–0.92) | ||||
Conclusions and Relevance
In the United States, there was a substantial decrease in mortality rates over time among children and adolescents initiating ESKD treatment with dialysis between 1990 and 2010. Further research is needed to determine the specific factors responsible for this decrease.
Individuals with end-stage kidney disease (ESKD) face a significantly shortened life expectancy.1 In no group of ESKD patients is the loss of potential years of life larger than in children and adolescents.2–4 Although transplant remains the treatment of choice to maximize survival, growth, and development, 75% of children with ESKD require treatment with dialysis prior to receiving a kidney transplant.1,5 Dialysis is therefore a lifesaving therapy for children with ESKD while they await transplant. Nevertheless, all-cause mortality rates in children receiving maintenance dialysis are at least 30 times higher than the general pediatric population, with even higher relative risks in very young children.2 The 2 most common causes of death in children with ESKD are cardiovascular disease and infection.2–4,6,7 There have been substantial improvements in the care of children with ESKD between 1990 and 2010. However, to our knowledge, it is not known if mortality has changed over time in the United States, particularly in recent years.2,6,7
The objective of this study was to determine if all-cause, cardiovascular, and infection-related mortality rates have changed between 1990 and 2010 among children and adolescents with ESKD initially treated with dialysis and if changes in mortality rates over time differed by age at treatment initiation. Based on a study in the Australian and New Zealand pediatric ESKD population2 and on the greater potential for improvement in younger children, we hypothesized that mortality rates have improved in the United States between 1990 and 2010 and that improvements have been greater for younger compared with older children.
METHODS
Data Source and Population
This was a retrospective cohort study of individuals recorded in the United States Renal Data System (USRDS) database who initiated ESKD treatment with dialysis for the first time at younger than 21 years of age between January 1, 1990, and December 31, 2010, and who were followed up until December 31, 2010. The USRDS includes virtually all children treated for ESKD in the United States (eMethods, available at http://www.jama.com).1,8 Because a comprehensive version of the death notification form came into use January 1, 1990, we chose this as the start date.1,4 Patients with a prior kidney transplant were excluded, as were patients receiving preemptive transplant. We divided the observation interval into 4 approximately equal periods to highlight how patient characteristics, cause of death, and the amount of missing data have changed over time.1,8 The institutional review board at The Children’s Hospital of Philadelphia approved the study.
Primary Exposure and Outcome Variables
The primary exposure was the calendar year of ESKD treatment initiation. The primary outcome was all-cause mortality. Cardiovascular and infection-related mortality were also considered as outcomes. Cause of death was determined from the USRDS death notification form (eMethods, Table 1). Deaths are reported to the USRDS via several different mechanisms, ensuring virtually complete capture.1
Table 1.
Patient Characteristics by Year of Treatment Initiation for End-Stage Kidney Disease
| Age <5 y at ESKD Treatment Initiation
|
Age ≥5 y at ESKD Treatment Initiation
|
|||||||
|---|---|---|---|---|---|---|---|---|
| 1990–1994 | 1995–1999 | 2000–2004 | 2005–2010 | 1990–1994 | 1995–1999 | 2000–2004 | 2005–2010 | |
| No. of patients | 692 | 734 | 845 | 1179 | 4368 | 4661 | 5103 | 5819 |
|
| ||||||||
| Age at first ESKD care, median (IQR), y | ||||||||
| All patients | 1.0 (0.2–2.6) | 1.1 (0.2–2.6) | 0.9 (0.1–2.4) | 0.6 (0.1–2.0) | 17.1 (13.5–19.3) | 17.0 (13.4–19.3) | 16.8 (13.3–19.3) | 17.1 (14.0–19.4) |
|
| ||||||||
| Initiating hemodialysis | 1.2 (0.4–3.2) | 1.6 (0.5–3.1) | 1.7 (0.3–3.5) | 0.5 (0.1–2.4) | 18.0 (15.3–19.7) | 18.0 (15.2–19.7) | 17.9 (14.8–19.7) | 18.0 (15.3–19.8) |
|
| ||||||||
| Initiating peritoneal dialysis | 0.8 (0.2–2.3) | 0.8 (0.1–2.4) | 0.7 (0.1–2.0) | 0.6 (0.1–1.9) | 15.0 (11.4–18.1) | 14.6 (11.2–17.6) | 14.1 (10.8–17.1) | 14.6 (11.1–17.1) |
|
| ||||||||
| Male sex, No. (%) | 432 (62.4) | 447 (60.9) | 522 (61.8) | 762 (64.6) | 2360 (54.0) | 2503 (53.7) | 2733 (53.6) | 3163 (54.4) |
|
| ||||||||
| Race, No. (%) | ||||||||
| White | 538 (77.8) | 541 (73.7) | 604 (71.5) | 858 (72.8) | 2720 (62.3) | 2776 (59.6) | 3043 (59.6) | 3830 (65.8) |
|
| ||||||||
| Black | 122 (17.6) | 136 (18.5) | 145 (17.2) | 183 (15.5) | 1342 (30.7) | 1523 (32.7) | 1533 (30.0) | 1473 (25.3) |
|
| ||||||||
| Other | 32 (4.6) | 57 (7.8) | 96 (11.4) | 138 (11.7) | 306 (7.0) | 362 (7.8) | 527 (10.3) | 516 (8.9) |
|
| ||||||||
| Primary kidney disease, No. (%)a | ||||||||
| CAKUT | 194 (28.0) | 286 (39.0) | 347 (41.1) | 447 (37.8) | 461 (10.6) | 815 (17.5) | 887 (17.4) | 1043 (17.9) |
|
| ||||||||
| Glomerulonephritis | 83 (12.0) | 42 (5.7) | 27 (3.2) | 35 (3.0) | 1557 (35.6) | 1627 (34.8) | 1655 (32.4) | 1723 (29.6) |
|
| ||||||||
| FSGS | 29 (4.2) | 44 (6.0) | 42 (5.0) | 59 (5.0) | 467 (10.7) | 669 (14.4) | 800 (15.7) | 866 (14.9) |
|
| ||||||||
| Other | 210 (30.3) | 311 (42.4) | 370 (43.7) | 497 (42.2) | 778 (17.8) | 811 (17.4) | 888 (17.4) | 1046 (18.0) |
|
| ||||||||
| Unknown | 37 (5.4) | 23 (3.1) | 53 (6.3) | 88 (7.5) | 556 (12.7) | 638 (13.7) | 826 (16.2) | 1025 (17.6) |
|
| ||||||||
| Missing | 139 (20.1) | 28 (3.8) | 6 (0.7) | 53 (4.5) | 549 (12.6) | 101 (2.2) | 47 (0.9) | 116 (2.0) |
|
| ||||||||
| Modality initiated, No. (%) | ||||||||
| Hemodialysis | 133 (19.2) | 142 (19.3) | 164 (19.4) | 290 (24.6) | 2540 (58.2) | 2898 (62.2) | 3358 (65.8) | 3869 (66.5) |
|
| ||||||||
| Peritoneal dialysis | 497 (71.8) | 538 (73.3) | 633 (74.9) | 793 (67.3) | 1567 (35.9) | 1586 (34.0) | 1511 (29.6) | 1555 (26.7) |
|
| ||||||||
| Missing | 62 (9.0) | 54 (7.4) | 48 (5.7) | 96 (8.1) | 261 (6.0) | 177 (3.8) | 234 (4.6) | 395 (6.8) |
|
| ||||||||
| Socioeconomic quartile, No. (%)a | ||||||||
| Lowest | 137 (19.8) | 175 (23.8) | 169 (20.0) | 197 (16.7) | 1118 (25.6) | 1136 (24.4) | 1253 (24.6) | 1378 (23.7) |
|
| ||||||||
| Mid-low | 133 (19.2) | 121 (16.5) | 159 (18.8) | 209 (17.7) | 798 (18.3) | 936 (20.1) | 959 (18.8) | 1150 (19.8) |
|
| ||||||||
| Mid-high | 167 (24.1) | 156 (21.3) | 194 (23.0) | 276 (23.4) | 1074 (24.6) | 1097 (23.5) | 1216 (23.8) | 1348 (23.2) |
|
| ||||||||
| Highest | 231 (33.4) | 249 (33.9) | 298 (35.3) | 401 (34.0) | 1171 (26.8) | 1272 (27.3) | 1454 (28.5) | 1655 (28.4) |
|
| ||||||||
| Missing | 24 (3.5) | 33 (4.5) | 25 (3.0) | 96 (8.1) | 207 (4.7) | 220 (4.7) | 221 (4.3) | 288 (5.0) |
|
| ||||||||
| Comorbidities, No. (%)a | ||||||||
| None | 210 (30.4) | 601 (81.9) | 711 (84.1) | 757 (64.2) | 1582 (36.2) | 3876 (83.2) | 4441 (87.0) | 4641 (79.8) |
|
| ||||||||
| ≥1 Comorbidity | 27 (3.9) | 101 (13.8) | 126 (14.9) | 367 (31.1) | 361 (8.3) | 655 (14.1) | 611 (12.0) | 1052 (18.1) |
|
| ||||||||
| Missing | 455 (65.8) | 32 (4.4) | 8 (1.0) | 55 (4.7) | 2425 (55.5) | 130 (2.8) | 51 (1.0) | 126 (2.2) |
|
| ||||||||
| Insurer, No. (%)a | ||||||||
| Public | 108 (15.6) | 267 (36.4) | 368 (43.6) | 567 (48.1) | 1024 (23.4) | 1509 (32.4) | 1809 (35.5) | 2382 (40.9) |
|
| ||||||||
| Private | 114 (16.5) | 367 (50.0) | 440 (52.1) | 533 (45.2) | 743 (17.0) | 2217 (47.6) | 2553 (50.0) | 2607 (44.8) |
|
| ||||||||
| No coverage | 11 (1.6) | 35 (4.8) | 28 (3.3) | 17 (1.4) | 128 (2.9) | 611 (13.1) | 683 (13.4) | 698 (12.0) |
|
| ||||||||
| Missing | 459 (66.3) | 65 (8.9) | 9 (1.1) | 62 (5.3) | 2473 (56.6) | 324 (7.0) | 58 (1.1) | 132 (2.3) |
|
| ||||||||
| eGFR, mL/min/1.73 m2 at initiation, median (IQR)a | 8.3 (6.2–11.4) | 7.3 (5.5–10.3) | 8.5 (6.2–12.0) | 9.8 (6.8–14.3) | 7.9 (5.8–10.7) | 6.6 (4.9–8.6) | 7.4 (5.4–9.8) | 8.2 (5.8–11.2) |
|
| ||||||||
| Missing, % | 71.8 | 43.1 | 36.0 | 36.6 | 62.8 | 12.7 | 3.6 | 4.4 |
|
| ||||||||
| Erythrocyte-stimulating agent use at initiation, No. (%)a | 122 (17.6) | 277 (37.7) | 333 (39.4) | 468 (39.7) | 731 (16.7) | 1330 (28.5) | 1772 (34.7) | 1857 (31.9) |
|
| ||||||||
| Missing, No. (%) | 492 (71.1) | 91 (12.4) | 10 (1.2) | 55 (4.7) | 2689 (61.6) | 456 (9.8) | 54 (1.1) | 126 (2.2) |
|
| ||||||||
| Dialysis time before first transplant, median (IQR), y | 1.3 (0.6–2.7) | 1.4 (0.7–2.6) | 1.5 (0.9–2.4) | 1.3 (0.7–2.1) | 1.3 (0.5–2.9) | 1.4 (0.6–3.0) | 1.5 (0.7–2.8) | 0.9 (0.5–1.6) |
|
| ||||||||
| No. of transplants before end of observation | 486 | 550 | 612 | 490 | 3467 | 3580 | 3668 | 2637 |
|
| ||||||||
| Cause of death during treatment with dialysis, No. (%)b | ||||||||
| Cardiovascular | 57 (31.5) | 46 (30.5) | 57 (31.5) | 52 (27.1) | 236 (36.3) | 260 (39.3) | 234 (37.9) | 122 (35.9) |
|
| ||||||||
| Infection | 25 (13.8) | 34 (22.5) | 40 (22.1) | 37 (19.3) | 104 (16.0) | 123 (18.6) | 95 (15.4) | 33 (9.7) |
|
| ||||||||
| Malignancy | 7 (3.9) | 9 (6.0) | 4 (2.2) | 7 (3.7) | 10 (1.5) | 8 (1.2) | 14 (2.3) | 15 (4.4) |
|
| ||||||||
| Hemorrhage | 2 (1.1) | 5 (3.3) | 4 (2.2) | 2 (1.0) | 15 (2.3) | 24 (3.6) | 25 (4.0) | 3 (0.9) |
|
| ||||||||
| Metabolic | 0 | 1 (0.7) | 2 (1.1) | 1 (0.5) | 18 (2.8) | 22 (3.3) | 8 (1.3) | 7 (2.1) |
|
| ||||||||
| Other | 9 (5.0) | 9 (6.0) | 11 (6.1) | 22 (11.5) | 39 (6.0) | 47 (7.1) | 61 (9.9) | 38 (11.2) |
|
| ||||||||
| Unknown | 24 (13.3) | 26 (17.2) | 35 (19.3) | 17 (8.9) | 87 (13.4) | 87 (13.2) | 65 (10.5) | 27 (7.9) |
|
| ||||||||
| Missing | 57 (31.5) | 21 (13.9) | 28 (15.5) | 54 (28.1) | 142 (21.8) | 90 (13.6) | 116 (18.8) | 95 (27.9) |
Abbreviations: CAKUT, congenital anomalies of the kidney and urinary tract; eGFR, estimated glomerular filtration rate; ESKD, end-stage kidney disease; FSGS, focal segmental glomerulosclerosis; IQR, interquartile range.
The changes in reporting of comorbidity, laboratory values, cause of death, medication use, and insurance status over time are shown in eTable 1.
Deaths were classified as cardiovascular if the cause of death was recorded as any of the following: acute myocardial infarction, pericarditis, cardiac tamponade, atherosclerotic heart disease, cardiomyopathy, cardiac arrhythmia, cardiac arrest (cause unknown), valvular heart disease, congestive heart failure, pulmonary embolus, or cerebrovascular accident including intracranial hemorrhage. Infection-related causes included septicemia, peritoneal access infection, peritonitis, viral hepatitis, viral infection, tuberculosis, AIDS, or infections of the central nervous system, heart, lungs, abdomen, and genitourinary systems.
Association Between Year of Initiation and Mortality Rate
We calculated mortality rates (deaths per 1000 person-years of observation) for each year of initiation and plotted the data to characterize the “shape” of the relationship. Although there was year-to-year variability in mortality, overall mortality rates decreased gradually and linearly with year of initiation; there were no clear “step” changes. Therefore, initiation year was treated as a continuous variable.
We generated Kaplan-Meier curves to illustrate differences in mortality among 4 approximately equal time periods within the 1990–2010 interval. We used Cox proportional hazards models to estimate the relative mortality (hazard ratios [HRs] and 95% CIs) associated with a 5-year increment in calendar year of ESKD treatment initiation from 1990–2010. Because we were interested in examining changes in mortality over calendar time during treatment with dialysis, observation was limited to the period of dialysis prior to a first transplant. Time zero was the day of dialysis initiation, and observation was censored at first transplant, death, or end of observation. We considered the possibility of different relationships between year of initiation and mortality in different age groups by including multiplicative year × age interaction terms in the models. Based on prior studies2,4,6 and the age categories used in the USRDS annual reports,1 age at initiation was divided into 2 categories (<5 years and ≥5 years). Proportionality of hazards was confirmed by examining plots of the data.
Although virtually all children treated for ESKD are considered eligible for transplantation, and therefore a problem with competing risks is unlikely, we performed sensitivity analyses in which the person-time of patients who were not censored because of transplant was weighted by the inverse of their probability of continuing to receive dialysis.9 The weights were determined by fitting a logistic regression model to estimate the probability of being censored because of transplant, given a particular profile of all the other covariates in the models. Such weighting creates a pseudo population without transplant censoring such that the weighted population is no longer a biased sample. In addition, we conducted analyses in which observation was not censored at transplant; these analyses include observation during treatment with dialysis as well as with transplant.
Models were initially adjusted for age at dialysis initiation (as a continuous variable within each stratum), sex, race, socioeconomic status (SES), primary kidney disease, and initial dialysis modality. Socioeconomic status was estimated using median household income by zip code and classified by quartile within the US Census data (1999).10 Subsequent models were also adjusted for insurance coverage and co-morbidity; these variables were associated with relatively greater amounts of missing data, especially in earlier time periods, so were added sequentially to the models. In addition, reporting of both insurance coverage and comorbidity has changed over time (eTable 1). Estimated glomerular filtration rate (eGFR; estimated from height and serum creatinine level within 3 months of dialysis initiation using the updated Schwartz formula11) and erythrocyte-stimulating agent (ESA) use at initiation were examined but were not included in the multivariable analyses because of very large amounts of missing data.
Imputation of Missing Covariates and Missing Cause of Death
For the all-cause mortality models, missing covariate values were imputed using multiple imputation methods12 and the joint distributions of all other variables in the models, including outcomes (eMethods). Similarly, for cardiovascular and infection-related mortality models, missing causes of death and missing covariates were imputed from the joint distributions of all other variables in the models, including nonmissing causes of death.12 We also report analyses including only cases with nonmissing cause of death.
Data analyses were performed using SAS version 9.2 (SAS Institute) and S-plus (version 6.1); a 2-sided P value <.05 was considered statistically significant.
RESULTS
Patient Characteristics and Causes of Death
We identified 23 401 children and adolescents who initiated ESKD treatment with dialysis at younger than 21 years from January 1990 until December 2010. Table 1 summarizes patient characteristics and causes of death by age stratum and over time. Age at ESKD treatment initiation decreased over time among those younger than 5 years. This decrease was more marked among those initiating hemodialysis than peritoneal dialysis. Although most younger children initiated peritoneal dialysis and older patients initiated hemodialysis, there was a slight increase in the use of he modialysis over time in both age groups.
Comorbidities were uncommon but appeared to increase slightly over time, especially in younger children. Predialysis care, as measured using eGFR and ESA use at initiation, did not appear to change over time. Cardiovascular disease and infection were the most frequently documented causes of death in both age strata and in all time periods. Cardiac arrest (cause unknown) was the most common cardiovascular cause of death in both younger (53%) and older (49%) individuals receiving dialysis.
Change in Mortality Risk Over Time
All-Cause Mortality
Crude mortality rates were higher in children younger than 5 years compared with those 5 years and older at initiation (Table 2). Mortality risk decreased progressively over calendar time for both those younger than 5 years and those 5 years and older at initiation. The Figure illustrates the survival, during dialysis treatment, of patients initiating ESKD care in 4 calendar time periods between 1990 and 2010 for each age stratum. Each 5-year increment in calendar year of dialysis initiation was associated with an adjusted HR of 0.80 (95% CI, 0.75–0.85) among children younger than 5 years at initiation and an HR of 0.88 (95% CI, 0.85–0.92) among those 5 years and older. The magnitude of this improvement was greater for younger children than older children; however, this difference did not reach statistical significance (interaction P=.10). The sensitivity analyses, including analyses without censoring at transplant and analyses in which inverse probability of censoring weighting was applied, returned qualitatively similar results (eFigure 1, eTable 2).
Table 2.
Crude All-Cause Mortality Rates and All-Cause Hazard Ratios for Mortality, Stratified by Age
| Age <5 y at ESKD Treatment Initiation | Age ≥5 y at ESKD Treatment Initiation | |
|---|---|---|
| 1990–1994 | ||
| Person-years of observation | 1613 | 14 595 |
|
| ||
| Deaths, No. | 181 | 651 |
|
| ||
| Crude all-cause mortality rate per 1000 person-years | 112.2 | 44.6 |
|
| ||
| 1995–1999 | ||
| Person-years of observation | 1550 | 15 749 |
|
| ||
| Deaths, No. | 151 | 661 |
|
| ||
| Crude all-cause mortality rate per 1000 person-years | 97.4 | 42.0 |
|
| ||
| 2000–2004 | ||
| Person-years of observation | 1670 | 15 350 |
|
| ||
| Deaths, No. | 181 | 618 |
|
| ||
| Crude all-cause mortality rate per 1000 person-years | 108.4 | 40.3 |
|
| ||
| 2005–2010 | ||
| Person-years of observation | 2303 | 13 104 |
|
| ||
| Deaths, No. | 192 | 340 |
|
| ||
| Crude all-cause mortality rate per 1000 person-years | 83.4 | 25.9 |
|
| ||
| Entire interval: 1990–2010 | ||
| Person-years of observation | 7136 | 58 799 |
|
| ||
| Deaths, No. | 705 | 2270 |
|
| ||
| Crude all-cause mortality rate per 1000 person-years | 98.8 | 38.6 |
|
| ||
| HR per 5-y increment in calendar year of ESKD treatment initiation (95% CI) | ||
| Unadjusted | 0.84 (0.78–0.89) | 0.87 (0.83–0.90) |
|
| ||
| Adjusted model 1a | 0.82 (0.77–0.87) | 0.88 (0.85–0.92) |
|
| ||
| Adjusted model 2 (model 1 plus insurance coverage) | 0.82 (0.77–0.87) | 0.89 (0.86–0.93) |
|
| ||
| Adjusted model 3 (model 2 plus comorbidity) | 0.80 (0.75–0.85) | 0.88 (0.85–0.92) |
Abbreviations: ESKD, end-stage kidney disease; HR, hazard ratio.
Model 1 adjusted for age at initiation, sex, race, primary renal disease, initial dialysis modality, and socioeconomic status quartile.
Figure. Kaplan-Meier Estimates of Survival for Children Initiating ESKD Treatment With Dialysis in Successive Time Periods.
Observation was censored at first transplant. The numbers of individuals still under observation, and still treated with dialysis, are indicated below the curves. ESKD indicates end-stage kidney disease.
The multivariable Cox models also permitted identification of factors independently associated with all-cause mortality (Table 3). Higher mortality risk was independently associated with glomerulonephritis and “other” primary renal disease (vs congenital anomalies of the kidney or urinary tract), presence of 1 or more comorbidity, and lower estimated SES in both age groups. In children younger than 5 years at initiation, younger age, unknown primary renal disease, and hemodialysis were associated with higher mortality risk. In children 5 years and older at initiation, female sex, black race (vs white), and public insurance (vs no insurance) were associated with higher mortality risk.
Table 3.
Associations Between Covariates and Mortality
| No. of Patients | Person-Years | No. of Deaths | Death Rate per 1000 Person-Years | Unadjusted HR (95% CI) | Adjusted HR (95% CI)a | |
|---|---|---|---|---|---|---|
| Age <5 y at ESKD Treatment Initiation | ||||||
| Age at initiation (per year) | 3450 | 7136 | 705 | 98.8 | 0.77 (0.73–0.82) | 0.75 (0.70–0.80) |
|
| ||||||
| Sex | ||||||
| Female | 1287 | 2655 | 293 | 110.4 | Reference | Reference |
|
| ||||||
| Male | 2163 | 4482 | 412 | 91.9 | 0.83 (0.72–0.97) | 0.86 (0.73–1.00) |
|
| ||||||
| Race | ||||||
| White | 2541 | 5036 | 480 | 95.3 | Reference | Reference |
|
| ||||||
| Black | 586 | 1408 | 150 | 106.5 | 1.24 (1.03–1.48) | 1.17 (0.96–1.41) |
|
| ||||||
| Other | 323 | 692 | 75 | 108.4 | 1.18 (0.93–1.50) | 1.22 (0.95–1.41) |
|
| ||||||
| Primary disease | ||||||
| CAKUT | 1274 | 2651 | 199 | 75.1 | Reference | Reference |
|
| ||||||
| Glomerulonephritis | 187 | 300 | 32 | 106.7 | 1.56 (1.13–2.14) | 1.49 (1.08–2.04) |
|
| ||||||
| FSGS | 174 | 321 | 12 | 37.4 | 0.93 (0.48–1.79) | 1.25 (0.78–2.00) |
|
| ||||||
| Other | 1388 | 2949 | 301 | 102.1 | 1.42 (1.17–1.71) | 1.35 (1.11–1.63) |
|
| ||||||
| Unknown | 201 | 390 | 80 | 205.1 | 2.68 (2.01–3.57) | 2.06 (1.56–2.73) |
|
| ||||||
| Dialysis modality | ||||||
| Hemodialysis | 729 | 1439 | 192 | 133.4 | Reference | Reference |
|
| ||||||
| Peritoneal dialysis | 2461 | 5408 | 418 | 77.3 | 0.62 (0.52–0.73) | 0.71 (0.58–0.85) |
|
| ||||||
| SES quartile | ||||||
| Lowest | 678 | 1497 | 176 | 117.6 | Reference | Reference |
|
| ||||||
| Mid-low | 622 | 1295 | 133 | 102.7 | 0.90 (0.69–1.18) | 0.90 (0.71–1.13) |
|
| ||||||
| Mid-high | 793 | 1565 | 148 | 94.6 | 0.79 (0.62–1.00) | 0.84 (0.67–1.05) |
|
| ||||||
| Highest | 1179 | 2392 | 212 | 88.6 | 0.77 (0.61–0.97) | 0.79 (0.64–0.98) |
|
| ||||||
| Comorbidity | ||||||
| None | 2279 | 4740 | 326 | 68.8 | Reference | Reference |
|
| ||||||
| ≥1 | 621 | 1230 | 189 | 153.7 | 1.81 (1.53–2.14) | 1.63 (1.10–1.34) |
|
| ||||||
| Insurance coverage | ||||||
| None | 91 | 176 | 20 | 113.6 | Reference | Reference |
|
| ||||||
| Public | 1310 | 2801 | 257 | 91.8 | 0.70 (0.49–1.01) | 0.81 (0.53–1.25) |
|
| ||||||
| Private | 1454 | 2904 | 231 | 79.5 | 0.60 (0.43–0.85) | 0.76 (0.47–1.23) |
|
| ||||||
| Age ≥5 y at ESKD Treatment Initiation | ||||||
| Age at initiation (per year) | 19 951 | 58 799 | 2270 | 38.6 | 1.01 (0.99–1.02) | 1.00 (0.99–1.02) |
|
| ||||||
| Sex | ||||||
| Female | 9192 | 27 529 | 1230 | 44.7 | Reference | Reference |
|
| ||||||
| Male | 10 759 | 31 269 | 1040 | 33.3 | 0.74 (0.68–0.81) | 0.79 (0.73–0.86) |
|
| ||||||
| Race | ||||||
| White | 12 369 | 30 923 | 1053 | 34.1 | Reference | Reference |
|
| ||||||
| Black | 5871 | 22 249 | 1021 | 45.9 | 1.32 (1.21–1.44) | 1.32 (1.21–1.45) |
|
| ||||||
| Other | 1711 | 5626 | 196 | 34.8 | 1.07 (0.87–1.19) | 1.09 (0.94–1.27) |
|
| ||||||
| Primary disease | ||||||
| CAKUT | 3206 | 7642 | 198 | 25.9 | Reference | Reference |
|
| ||||||
| Glomerulonephritis | 6562 | 20 328 | 822 | 40.4 | 1.48 (1.27–1.72) | 1.33 (1.13–1.56) |
|
| ||||||
| FSGS | 2802 | 8015 | 190 | 23.7 | 1.01 (0.84–1.22) | 0.91 (0.75–1.10) |
|
| ||||||
| Other | 3523 | 9236 | 578 | 62.6 | 2.26 (1.93–2.64) | 1.92 (1.64–2.25) |
|
| ||||||
| Unknown | 3045 | 10 790 | 290 | 26.9 | 1.00 (0.84–1.20) | 0.92 (0.76–1.11) |
|
| ||||||
| Dialysis modality | ||||||
| Hemodialysis | 12 665 | 41 308 | 1566 | 37.9 | Reference | Reference |
|
| ||||||
| Peritoneal dialysis | 6219 | 15 937 | 537 | 33.7 | 0.93 (0.84–1.02) | 0.96 (0.87–1.07) |
|
| ||||||
| SES quartile | ||||||
| Lowest | 4885 | 16 928 | 738 | 43.6 | Reference | Reference |
|
| ||||||
| Mid-low | 3843 | 11 603 | 446 | 38.4 | 0.91 (0.81–1.02) | 0.90 (0.80–1.01) |
|
| ||||||
| Mid-high | 4735 | 13 770 | 504 | 36.6 | 0.86 (0.77–0.97) | 0.88 (0.78–1.00) |
|
| ||||||
| Highest | 5552 | 13 745 | 471 | 34.3 | 0.82 (0.73–0.92) | 0.86 (0.75–0.98) |
|
| ||||||
| Comorbidity | ||||||
| None | 14 540 | 39 866 | 1105 | 27.7 | Reference | Reference |
|
| ||||||
| ≥1 | 2679 | 8051 | 553 | 68.7 | 1.98 (1.78–2.21) | 1.84 (1.65–2.07) |
|
| ||||||
| Insurance coverage | ||||||
| None | 2120 | 7660 | 202 | 26.4 | Reference | Reference |
|
| ||||||
| Public | 6724 | 20 466 | 815 | 39.8 | 1.37 (1.17–1.60) | 1.25 (1.07–1.46) |
|
| ||||||
| Private | 8120 | 18 802 | 595 | 31.6 | 1.12 (0.98–1.29) | 1.12 (0.96–1.30) |
Abbreviations: CAKUT, congenital anomalies of the kidney and urinary tract; ESKD, end-stage kidney disease; FSGS, focal segmental glomerulosclerosis; HR, hazard ratio; SES, socioeconomic status.
Adjusted for age at initiation, sex, race, primary renal disease, initial dialysis modality, SES quartile, comorbidity, and insurance status (model 3).
Cardiovascular and Infection-Related Mortality
Table 4 shows crude cardiovascular and infection-related mortality rates for both age groups as well as the HRs for mortality associated with a 5-year increment in calendar year of ESKD treatment initiation. For cardiovascular mortality, each 5-year increment in calendar year of initiation of dialysis was associated with an adjusted HR of 0.54 (95% CI, 0.47–0.63) among children younger than 5 years at initiation and an HR of 0.66 (95% CI, 0.61–0.70) among those 5 years and older. For infection-related mortality, each 5-year increment in calendar year of initiation of dialysis was associated with an adjusted HR of 0.64 (95% CI, 0.52–0.79) among children younger than 5 years at initiation and an HR of 0.59 (95% CI, 0.53–0.65) among those 5 years and older. The sensitivity analyses, including analyses without censoring at transplant and analyses including only cases with non-missing cause of death, were similar (eTable 3).
Table 4.
Cardiovascular and Infection-Related Mortality by Age at Treatment Initiation for End-Stage Kidney Disease
| Cardiovascular Mortality | Infection-Related Mortality | |||
|---|---|---|---|---|
| Age <5 y | Age ≥5 y | Age 5 y | Age ≥5 y | |
| 1990–1994 | ||||
| Person-years of observation | 1613 | 14 595 | 1613 | 14 595 |
| Deaths, No. | 57 | 236 | 25 | 104 |
| Crude mortality rate per 1000 person-years | 35.3 | 16.2 | 15.5 | 7.1 |
| 1995–1999 | ||||
| Person-years of observation | 1550 | 15 749 | 1550 | 15 749 |
| Deaths, No. | 46 | 260 | 34 | 123 |
| Crude mortality rate per 1000 person-years | 29.7 | 16.5 | 21.9 | 7.8 |
| 2000–2004 | ||||
| Person-years of observation | 1670 | 15 350 | 1670 | 15 350 |
| Deaths, No. | 57 | 234 | 40 | 95 |
| Crude mortality rate per 1000 person-years | 34.1 | 15.2 | 24.0 | 6.2 |
| 2005–2010 | ||||
| Person-years of observation | 2303 | 13 104 | 2303 | 13 104 |
| Deaths, No. | 52 | 122 | 37 | 33 |
| Crude mortality rate per 1000 person-years | 22.6 | 9.3 | 16.0 | 2.5 |
| Entire interval: 1990–2010 | ||||
| Person-years of observation | 7136 | 58 799 | 7136 | 58 799 |
| Deaths, No. | 246 | 852 | 136 | 355 |
| Crude mortality rate per 1000 person-years | 34.5 | 14.5 | 19.1 | 6.0 |
| HR per 5-y increment in calendar year of ESKD treatment initiation (95% CI) | ||||
| Unadjusted | 0.80 (0.71–0.90) | 0.90 (0.85–0.96) | 0.88 (0.75–1.04) | 0.82 (0.74–0.90) |
| Adjusted model 1a | 0.77 (0.69–0.86) | 0.92 (0.86–0.98) | 0.84 (0.73–0.96) | 0.83 (0.76–0.91) |
| Adjusted model 2 (model 1 plus insurance coverage) | 0.77 (0.67–0.87) | 0.92 (0.87–0.98) | 0.84 (0.71–0.99) | 0.84 (0.76–0.92) |
| Adjusted model 3 (model 2 plus comorbidity) | 0.54 (0.47–0.63) | 0.66 (0.61–0.70) | 0.64 (0.52–0.79) | 0.59 (0.53–0.65) |
Abbreviations: ESKD, end-stage kidney disease; HR, hazard ratio.
Adjusted for age at initiation, sex, race, primary renal disease, initial dialysis modality, and socioeconomic status quartile.
DISCUSSION
To our knowledge, this study represents the largest cohort of pediatric ESKD patients ever examined, including more than 20 000 individuals who initiated dialysis over the course of 2 decades and were followed up until the end of 2010. We demonstrate that mortality rates for children and adolescents being treated for ESKD with dialysis improved significantly between 1990 and 2010. Although gains in survival appeared slightly greater for children younger than 5 years at initiation of ESKD care, the interaction between age and year of dialysis initiation was not statistically significant.
Our findings expand on the results of prior work. A study of 1634 Australian and New Zealand children (<20 years) initiating treatment for ESKD from 1963–2002 reported significantly lower mortality among those initiating dialysis after 1983 (compared with before 1983) but not after 1993 (compared with earlier years).2 Although the sample was small, the youngest children appeared to experience the greatest improvements in survival over time. A Dutch study of 381 patients younger than 15 years at ESKD initiation found lower mortality among those starting dialysis after 1982 (vs before 1982).7 In contrast, no significant changes in all-cause or cardiac mortality were seen between 1991 and 1996 in a USRDS study of 1454 incident dialysis patients (<20 years).6 These prior studies all identified children younger than 5 to 6 years as having the highest risk for death.2,6,7
The association observed between mortality and age in children with ESKD mirrors the association observed in the general pediatric population: mortality risk increases with decreasing age less than 5 years and increases with increasing age beyond about 10 years.13 However, the causes of death in children with ESKD are very different from those in the general population. Whereas accidents and homicides are the most common causes of death in the general population of children and adolescents,14,15 cardiovascular disease and infection are the 2 leading causes of death in those with ESKD.2–4,7,16
Numerous factors may have contributed to the observed reductions in mortality risk over time. Improved predialysis care, advances in dialysis technology, and greater experience of clinicians may each have played a role.17–20 We were limited in our ability to examine predialysis care by the information available in this comprehensive US ESKD registry. Large proportions of patients—particularly younger children and those initiating treatment in earlier years—had data missing for ESA use prior to and for eGFR at initiation of dialysis. Therefore, we can only speculate on which specific factors may be responsible for the observed improvements over time in mortality. eFigure 2 highlights advances in dialysis care over time, including regulatory approval of new medications,21 publication of clinical practice guidelines, and improved technology for performing dialysis.17,22–25 Although most improvements occurred before 2000, it may take several years to implement new technology and guidelines,26 and year of implementation likely varies across centers.
If there were greater improvements in all-cause mortality rates for younger compared with older children, these may be related to changes specifically targeting small children, including dialysis equipment geared to small body size and age-specific surgical techniques for placing vascular access for dialysis.17,27,28 It is possible the increasing experience of clinicians treating small children, combined with the implementation of new technologies, may also have played a role.
The broader implementation of clinical practice guidelines29 may explain some of the improvements in cause-specific mortality. For example, the use of downward-facing, double-cuffed, swan neck peritoneal dialysis catheters increased from 5% (1992–1995) to 24% (after 2002), potentially decreasing the risk of peritonitis, especially in young children.17 Strict infection control practices may have decreased hemodialysis catheter infections.30 Pneumococcal and influenza vaccination, which are associated with decreased mortality in adults with ESKD,31 have been increasingly used over the past decade in children with ESKD.1 Improvement in anemia care may have contributed to decreased cardiovascular risk. Anemia is associated with a higher risk of death in children initiating dialysis.32,33 It is therefore worth noting that the mean hemoglobin levels in children with ESKD increased from 9 g/dL in 1991 to more than 11 g/dL in 2007, paralleling an increase in weekly ESA dosing.1 However, the ideal hemoglobin target for adults treated with ESAs for ESKD has yet to be established, and there have been no trials on the effect of ESA therapy on outcomes in children.
Because time receiving dialysis is so brief in children (median <2 years) compared with adults, it is important to recognize that even the dramatic improvements in mortality risk over time that we observed may not translate into major improvements in overall mortality in children with ESKD. Waiting times for transplant have fallen for older children, primarily since 2005.34 Therefore, time of exposure to dialysis and its associated higher risk of death compared with transplant5 have decreased. However, for individuals requiring prolonged treatment with dialysis, including infants who must grow large enough to safely receive transplants, the reductions in mortality risk that have occurred over time may have a notable effect.
The improvements observed in cardiovascular and infection-related mortality rates over time during treatment with dialysis appeared to be larger than the improvements in all-cause mortality. This suggests that mortality from other causes may have increased over calendar time. However, these results must be interpreted with caution. Substantial numbers of patients had data missing for cause of death. Some covariates were also missing in large proportions, particularly in earlier years. In addition, there were changes over time in reporting of both causes of death and comorbidities. We used multiple imputation to deal with missing covariates and missing causes of death. Although this method is far less likely to introduce bias than including only cases with complete data,12 it does have limitations, which may be amplified as the proportion of patients with missing data increases. Imputation assumes that one can accurately predict the value of a variable from the other variables in the model. Where both the outcome (cause of death) and covariates are missing in the same patients, predictive capacity may be weak. The marked change in the magnitude of the HRs when comorbidity was included in the models, compared with when it was excluded, suggests either that comorbidity increased substantially over time or, more likely, that the imputation broke down on inclusion of comorbidity.
The HRs associated with some of the covariates included in our models suggest that certain children and adolescents undergoing dialysis may be at a higher risk of death than others. These HRs must be interpreted cautiously; our study was not designed to identify risk factors for death. Furthermore, the wide confidence intervals for some of these estimates suggest that power was limited. Some of the factors we identified, including black race, lower SES, and presence of comorbidity, have previously been shown to be associated with higher mortality risk.4–6,35 A higher independent risk of death in girls and young women, compared with boys and young men, was observed previously5 and deserves further investigation. The lower risk of death that we observed in children younger than 5 years initiating peritoneal dialysis, compared with hemodialysis, has also been described by others.1 It is important to recognize that we could only assess the initial dialysis modality. Up to 20% of pediatric dialysis patients switch modalities, often within the first few months of treatment.36 Prior analyses in which dialysis modality was treated as a time-dependent variable found higher mortality rates associated with peritoneal dialysis.2 We observed a higher risk of death in older children with public insurance compared with no insurance. This finding may simply reflect problems imputing insurance status, which was frequently missing in early years.
Our analysis has several strengths, including a large sample size and virtually complete capture of deaths. However, we acknowledge limitations. We cannot exclude the possibility that residual confounding by variables not captured in the USRDS contributed to the associations observed.37 Changes in the reporting of some variables between 1990 and 2010 may also have limited our ability to completely adjust for all potential confounders. In particular, our ability to completely adjust for insurance coverage and comorbidities was limited by large amounts of missing data in early years. Selection bias may have played a role in the observed improvements in mortality rates over time if physicians tended to not offer dialysis to sicker patients in more recent compared with more remote years. However, the stable or increasing prevalence of comorbidities over time argues against such bias. Our ability to adjust for SES was limited by the measure used to reflect SES: median household income by zip code provides a fairly rough estimate of individual-level SES.
We also acknowledge that cardiovascular mortality rates were strongly influenced by the inclusion of the code “cardiac arrest (cause unknown).” This is a particularly problematic code because it may include arrhythmias due to hyperkalemia, embolic events, or other noncardiac conditions. Including “cardiac arrest (cause unknown)” as a cardiovascular death may result in overestimation of cardiovascular mortality rates. However, prior studies in both the United States and elsewhere reported similar cardiovascular mortality rates in the pediatric dialysis population.2,7
Almost all children initiating ESKD treatment are considered eligible for transplant.2,5 However, most will require dialysis during their lifetime, either before transplant or after allograft loss.5 In the United States, there was a significant decrease in mortality rates over time among children and adolescents initiating ESKD treatment with dialysis between 1990 and 2010. Further research is needed to determine the specific factors responsible for this decrease.
Acknowledgments
Role of the Sponsor: The funding agencies had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript.
Funding/Support: Dr Mitsnefes is supported by the National Institutes of Health (R01DK07695, K24DK090070, and U01DK-03-012). Dr Laskin is supported by a Career Development Award in Comparative Effectiveness Research (1KM1CA156715-01). Dr Foster, a member of the McGill University Health Centre Research Institute (supported in part by the Fonds de la recherche en santé du Québec [FRSQ]), received salary support from the FRSQ. This work was supported in part by Health Resources and Services Administration contract 234-2005-370011C. Data were supplied by the USRDS.
Footnotes
Disclaimer: The content is the sole responsibility of the authors alone and does not necessarily reflect the views or policies of the Department of Health and Human Services or the United States government.
Author Contributions: Dr Foster had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Mitsnefes and Laskin contributede qually to the study.
Study concept and design: Mitsnefes, Laskin, Foster.
Acquisition of data: Mitsnefes, Laskin.
Analysis and interpretation of data: Mitsnefes, Laskin, Dahhou, Zhang, Foster.
Drafting of the manuscript: Mitsnefes, Laskin, Foster. Critical revision of the manuscript for important intellectual content: Mitsnefes, Laskin, Dahhou, Zhang, Foster.
Statistical analysis: Dahhou, Zhang, Foster. Obtained funding: Mitsnefes, Foster.
Administrative, technical, or material support: Laskin, Foster.
Study supervision: Mitsnefes, Foster.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Laskin is a site principal investigator for studies sponsored by Amgen and Genzyme but does not receive any funding from these companies. No other disclosures were reported.
References
- 1.2012 Annual data report: atlas of chronic kidney disease and end-stage renal disease in the United States. United States Renal Data System (USRDS); [Accessed April 12, 2013]. http://www.usrds.org/adr.aspx. [Google Scholar]
- 2.McDonald SP, Craig JC. Australian and New Zealand Paediatric Nephrology Association. Long-term survival of children with end-stage renal disease. N Engl J Med. 2004;350(26):2654–2662. doi: 10.1056/NEJMoa031643. [DOI] [PubMed] [Google Scholar]
- 3.Oh J, Wunsch R, Turzer M, et al. Advanced coronary and carotid arteriopathy in young adults with childhood-onset chronic renal failure. Circulation. 2002;106(1):100–105. doi: 10.1161/01.cir.0000020222.63035.c0. [DOI] [PubMed] [Google Scholar]
- 4.Parekh RS, Carroll CE, Wolfe RA, Port FK. Cardio vascular mortalityin children and young adults with end-stage kidney disease. J Pediatr. 2002;141(2):191–197. doi: 10.1067/mpd.2002.125910. [DOI] [PubMed] [Google Scholar]
- 5.Foster BJ, Dahhou M, Zhang X, Platt RW, Hanley JA. Change in mortality risk over time in young kidney transplant recipients. Am J Transplant. 2011;11(11):2432–2442. doi: 10.1111/j.1600-6143.2011.03691.x. [DOI] [PubMed] [Google Scholar]
- 6.Chavers BM, Li S, Collins AJ, Herzog CA. Cardiovascular disease in pediatric chronic dialysis patients. Kidney Int. 2002;62(2):648–653. doi: 10.1046/j.1523-1755.2002.00472.x. [DOI] [PubMed] [Google Scholar]
- 7.Groothoff JW, Gruppen MP, Offringa M, et al. Mortality and causes of death of end-stage renal disease in children: a Dutch cohort study. Kidney Int. 2002;61(2):621–629. doi: 10.1046/j.1523-1755.2002.00156.x. [DOI] [PubMed] [Google Scholar]
- 8.Foley RN, Collins AJ. The USRDS: what you need to know about what it can and can’t tell us about ESRD [published online ahead of print November 2, 2012] Clin J Am Soc Nephrol. doi: 10.2215/CJN.06840712. [DOI] [PubMed] [Google Scholar]
- 9.Hernán MA, Brumback B, Robins JM. Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men. Epidemiology. 2000;11(5):561–570. doi: 10.1097/00001648-200009000-00012. [DOI] [PubMed] [Google Scholar]
- 10.Woodward RS, Page TF, Soares R, Schnitzler MA, Lentine KL, Brennan DC. Income-related disparities in kidney transplant graft failures are eliminated by Medicare’s immunosuppression coverage. Am J Transplant. 2008;8(12):2636–2646. doi: 10.1111/j.1600-6143.2008.02422.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Schwartz GJ, Muñoz A, Schneider MF, et al. New equations to estimate GFR in children with CKD. J Am Soc Nephrol. 2009;20(3):629–637. doi: 10.1681/ASN.2008030287. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Schafer JL. Multiple imputation: a primer. Stat Methods Med Res. 1999;8(1):3–15. doi: 10.1177/096228029900800102. [DOI] [PubMed] [Google Scholar]
- 13.Hoyert DL, Xu JQ. Deaths: preliminary data for 2011. Natl Vital Stat Rep. 2012;61(6) [PubMed] [Google Scholar]
- 14.Mitsnefes MM. Cardiovascular disease in children with chronic kidney disease. J Am Soc Nephrol. 2012;23(4):578–585. doi: 10.1681/ASN.2011111115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mathews TJ, Miniño AM, Osterman MJ, Strobino DM, Guyer B. Annual summary of vital statistics: 2008. Pediatrics. 2011;127(1):146–157. doi: 10.1542/peds.2010-3175. [DOI] [PubMed] [Google Scholar]
- 16.Chavers BM, Solid CA, Gilbertson DT, Collins AJ. Infection-related hospitalization rates in pediatric versus adult patients with end-stage renal disease in the United States. J Am Soc Nephrol. 2007;18(3):952–959. doi: 10.1681/ASN.2006040406. [DOI] [PubMed] [Google Scholar]
- 17.Warady BA, Schaefer F, Alexander SR, editors. Pediatric Dialysis. New York, NY: Springer; 2012. [Google Scholar]
- 18.Sebestyen JF, Warady BA. Advances in pediatric renal replacement therapy. Adv Chronic Kidney Dis. 2011;18(5):376–383. doi: 10.1053/j.ackd.2011.07.003. [DOI] [PubMed] [Google Scholar]
- 19.Wedekin M, Ehrich JH, Offner G, Pape L. Renal replacement therapy in infants with chronic renal failure in the first year of life. Clin J Am Soc Nephrol. 2010;5(1):18–23. doi: 10.2215/CJN.03670609. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Vidal E, Edefonti A, Murer L, et al. Italian Registry of Paediatric Chronic Dialysis. Peritoneal dialysis in infants: the experience of the Italian Registry of Paediatric Chronic Dialysis. Nephrol Dial Transplant. 2012;27(1):388–395. doi: 10.1093/ndt/gfr322. [DOI] [PubMed] [Google Scholar]
- 21.Drug approvals and databases: Drugs@FDA Database. US Food and Drug Administration; [Accessed February 1, 2013]. http://www.fda.gov/Drugs/InformationOnDrugs/ucm135821.htm. [Google Scholar]
- 22.Tetta C, Roy T, Gatti E, Cerutti S. The rise of hemodialysis machines: new technologies in minimizing cardiovascular complications. Expert Rev Cardiovasc Ther. 2011;9(2):155–164. doi: 10.1586/erc.10.194. [DOI] [PubMed] [Google Scholar]
- 23.Dell’ Aquila R, Rodighiero MP, Spanó E, et al. Advances in the technology of automated, tidal, and continuous flow peritoneal dialysis. Perit Dial Int. 2007;27(suppl 2):S130–S137. [PubMed] [Google Scholar]
- 24.Ing TS, Rahman M, Kjellstrand CM, editors. Dialysis History, Development, and Promise. Singapore: World Scientific Publishing; 2012. [Google Scholar]
- 25.de Vries JP, Kouw PM, van der Meer NJ, et al. Non-invasive monitoring of blood volume during hemodialysis: its relation with postdialytic dry weight. Kidney Int. 1993;44(4):851–854. doi: 10.1038/ki.1993.321. [DOI] [PubMed] [Google Scholar]
- 26.Balas EA, Boren SA. Managing clinical knowledge for health care improvement. In: Bemmel J, McCray AT, editors. Yearbook of Medical Informatics 2000: Patient-Centered Systems. Institute for Healthcare Improvement; [Accessed February 8, 2013]. http://www.ihi.org/knowledge/Pages/Publications/Managingclinicalknowledgeforhealthcareimprovement.aspx. [PubMed] [Google Scholar]
- 27.Fischbach M, Edefonti A, Schróder C, Watson A. European Pediatric Dialysis Working Group. Hemodialysis in children: general practical guidelines. Pediatr Nephrol. 2005;20(8):1054–1066. doi: 10.1007/s00467-005-1876-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Verrina E, Cappelli V, Perfumo F. Selection of modalities, prescription, and technical issues in children on peritoneal dialysis. Pediatr Nephrol. 2009;24 (8):1453–1464. doi: 10.1007/s00467-008-0848-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.National Kidney Foundation. KDOQI clinical practice guidelines and clinical practice recommendations for 2006 updates: hemodialysis adequacy, peritoneal dialysis adequacy and vascular access. Am J Kidney Dis. 2006;48(suppl 1):S1–S322. [Google Scholar]
- 30.Eisenstein I, Tarabeih M, Magen D, et al. Low infection rates and prolonged survival times of hemodialysis catheters in infants and children. Clin J Am Soc Nephrol. 2011;6(4):793–798. doi: 10.2215/CJN.04810610. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Bond TC, Spaulding AC, Krisher J, McClellan W. Mortality of dialysis patients according to influenza and pneumococcal vaccination status. Am J Kidney Dis. 2012;60(6):959–965. doi: 10.1053/j.ajkd.2012.04.018. [DOI] [PubMed] [Google Scholar]
- 32.Warady BA, Ho M. Morbidity and mortality in children with anemia at initiation of dialysis. Pediatr Nephrol. 2003;18(10):1055–1062. doi: 10.1007/s00467-003-1214-1. [DOI] [PubMed] [Google Scholar]
- 33.Borzych-Duzalka D, Bilginer Y, Ha IS, et al. International Pediatric Peritoneal Dialysis Network (IPPN) Registry. Management of anemia in children receiving chronic peritoneal dialysis. J Am Soc Nephrol. 2013;24(4):665–676. doi: 10.1681/ASN.2012050433. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Matas AJ, Smith JM, Skeans MA, et al. OPTN/SRTR 2011 annual data report: kidney. Am J Transplant. 2013;13(suppl 1):11–46. doi: 10.1111/ajt.12019. [DOI] [PubMed] [Google Scholar]
- 35.Kucirka LM, Grams ME, Lessler J, et al. Association of race and age with survival among patients undergoing dialysis. JAMA. 2011;306(6):620–626. doi: 10.1001/jama.2011.1127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Leonard MB, Donaldson LA, Ho M, Geary DF. A prospective cohort study of incident maintenance dialysis in children: an NAPRTC study. Kidney Int. 2003;63(2):744–755. doi: 10.1046/j.1523-1755.2003.00788.x. [DOI] [PubMed] [Google Scholar]
- 37.Miskulin DC, Meyer KB, Martin AA, et al. Choices for Healthy Outcomes in Caring for End-Stage Renal Disease (CHOICE) Study. Comorbidity and its change predict survival in incident dialysis patients. Am J Kidney Dis. 2003;41(1):149–161. doi: 10.1053/ajkd.2003.50034. [DOI] [PubMed] [Google Scholar]