Introduction
Children with ESRD receiving chronic dialysis are hospitalized approximately 1–2 times per year (1,2), a rate 40–70 times higher than that of the general pediatric population (3) in the United States. The admission rate for pediatric dialysis patients has been increasing over the last decade, and about 35% of patients with ESRD age 0–19 years are readmitted to the hospital within 30 days of discharge (1). Hospitalizations are a burden for patients, caregivers, and the health care system. Admissions represent greater illness acuity and have been associated with higher mortality among adults undergoing chronic dialysis (4). The Centers for Medicare & Medicaid Services is considering adding a 30-day readmission measure to the ESRD Quality Incentive Program (5) and penalizing dialysis units with high readmission rates.
Previous studies examining readmissions in children do not provide a detailed view of the dialysis population (6,7). Our objective was to examine hospital readmissions among a cohort of children receiving chronic dialysis to identify factors associated with higher rates of 30-day readmission in this vulnerable population.
Materials and Methods
Data Source
The Pediatric Hospital Information System (PHIS) contains data from 43 free-standing children’s hospitals in the United States. The database contains information on >3.3 million inpatient admissions. Participating hospitals report demographic, clinical, and charge information for each hospitalization. Patients have unique medical record numbers, and each admission has a unique discharge identification allowing for tracking of patients over time. The data are sufficiently deidentified, and the study was considered exempt from institutional review board review.
Cohort Selection
We included index admissions of patients age 3 months to 17 years to PHIS hospitals with complete clinical and charge data reporting (n=38 of 43 hospitals) occurring between January 1, 2006, and November 30, 2010, and followed patients until December 31, 2010, to allow for 30 days of follow-up. To capture chronic dialysis patients, admissions with both an associated International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) procedure code for hemodialysis (HD) or peritoneal dialysis (PD) and an ESRD-identifying diagnosis (Table 1) were selected. Admissions without an ESRD-identifying diagnosis (n=729) were excluded because these may represent patients receiving acute inpatient dialysis and not those with ESRD. We included index hospitalizations with any patient disposition, but eliminated any admissions during which a patient died. Readmissions that met inclusion criteria were analyzed as an index admission because this allowed for the most complete evaluation of the outcome measures.
;)
Table 1: Diagnoses and procedures used to identify patients as having ESRD
Outcome Measures
The unit of analysis was the hospital admission. The primary outcome of interest was admission for any cause within 30 days of discharge from the index hospitalization (all-cause readmission), and the secondary outcome was admission for a principal diagnosis similar to that of the index admission within 30 days of discharge (similar-cause readmission). Readmission was considered a similar-cause readmission if it fell within the same principal diagnosis group as the index admission. (see below and Supplemental Table 1 for description of the principal diagnosis groups). In a separate analysis, we compared the characteristics of patients with three or fewer admissions to those with four or more admissions during the study period.
Admitting Diagnoses
Principal diagnoses were characterized into seven groups to test for associations between cause of admission and the odds of readmission (Supplemental Table 1 lists ICD-9-CM diagnoses included in each group): (1) peritonitis and PD catheter-related infection or complication, (2) HD access infection or complication, (3) cardiovascular admission, (4) electrolyte and fluid disorders, (5) complication of transplanted kidney, (6) other infection, and (7) all other diagnoses. The rate of 30-day readmission for each diagnosis group was examined, and association between diagnosis group and readmission was tested. The groups were used to test our secondary outcome of readmission for similar principal diagnosis within 30 days of discharge.
Explanatory Variables
Multiple variables were tested for association with the odds for 30-day readmission. These included patient demographic characteristics and admission-level factors. Age was divided into groups of younger than and older than 2 years at the time of admission because prior studies have shown that children younger than 2 years of age have a higher likelihood of readmission (8). Race was divided into black versus all other. Primary payer was divided into commercial, Medicaid, and Medicare.
Patient comorbid conditions were ascertained from secondary diagnosis codes and medication charges listed for each admission (Supplemental Table 2 lists diagnoses and medications used to define comorbid conditions). Comorbid conditions were categorized into groups (anemia, electrolyte abnormality, nutritional deficiency, psychiatric diagnosis, bone mineral disease, and hypertension), and the total number of comorbid conditions for each hospitalization was calculated by adding these groups at the admission level.
Dialysis modality was defined by the ICD-9-CM procedure code for the modality of dialysis performed during the hospitalization. In the case of a patient admitted for peritonitis or complications of a PD catheter, the long-term modality was presumed to be PD even if a patient received HD during the hospitalization.
The underlying renal diagnosis, or the presumed cause of ESRD, was identified for a patient by the ICD-9-CM admission, principal, or secondary diagnoses codes related to each admission and divided into glomerular or congenital/obstructive disease (Supplemental Table 3 lists ICD-9-CM codes and diagnoses used to define underlying renal diagnosis). When an underlying diagnosis was identified for a patient, it was applied to all the patient’s admissions. In the case of a patient with ICD-9-CM codes consistent with both glomerular and obstructive/congenital disease (n=83), the patient was categorized as having glomerular disease because we assumed the patient’s admission pattern would more closely resemble that of a patient with glomerular disease.
Statistical Analyses
Descriptive statistics for continuous variables were reported as medians and interquartile ranges (IQRs), and categorical variables were reported as proportions. A chi-square test was used to compare patients with three or fewer admissions to those with four or more admissions during the study period. The association between each explanatory variable and the primary outcome of 30-day readmission was tested using mixed-effects logistic regression clustering by patient within hospital to account for repeated admissions across multiple centers. All explanatory variables were tested initially in univariate analysis. Variables that were found to be statistically significant in univariate analysis along with dialysis modality were included in a multivariate model. All analyses were performed using Stata statistical software, version 12.0 (Stata Corp., College Station, TX).
Results
Descriptive Statistics
We identified 5112 index admissions among 1857 children with ESRD receiving chronic dialysis across 38 hospitals from January 1, 2006, until November 30, 2010. Table 2 lists population characteristics. Of the 5112 admissions, 2723 (53%) were of male patients, the median age was 12 years (IQR, 4–15 years), and 1628 (32%) were of black patients. An underlying diagnosis was identified in 3307 (65%) of admissions, and it was glomerular disease in 1988 (60%). HD patients made up 2518 of the admissions (49%). The median length of stay was 4 days (IQR, 2–9 days). The primary payer was commercial insurance, Medicare, and Medicaid in 20%, 34%, and 35% of admissions, respectively.
Table 2: Characteristics of index hospitalization, 30-day readmission, and unadjusted odds ratio for all-cause 30-day readmission
Principal diagnoses for the index admissions are presented in Supplemental Table 1. There were 526 (10.3%) admissions due to HD access infection or complication, 756 (14.8%) due to peritonitis and PD catheter-related infection or complication, and 610 (11.9%) due to other infection (of which 33.4% were viral). There were 823 (16.1%) admissions for cardiovascular diagnoses, 296 (5.6%) for electrolyte and fluid disorders, and 203 (4.0%) for complication of transplanted kidney. All other diagnoses made up 1898 (37.2%) of index admissions, and, of these, 1149 (60.8%) had a nonspecific principal diagnosis of ESRD, CKD, or a renal disease.
Of the 1857 children, 840 (45%) had a single admission during the study period, 372 (20%) patients had two admissions, and 216 (12%) had three admissions. These 1428 patients (77%) accounted for 44% of admissions. The remaining 429 (23%) patients had four or more admissions during the study period and accounted for 56% of admissions. Table 3 compares the characteristics of patients with three or fewer admissions to those with four or more admissions. Patients with four or more admissions, analyzed at the patient level, were more likely to be younger (P=0.001), to be black (P<0.001), to have public insurance (P<0.001), and to have comorbid conditions (P<0.001).
Table 3: Characteristics of patients with three or fewer admission or four or more admissions during the study period
30-Day Readmission
A total of 1282 (25%) index admissions were followed by a readmission within 30 days of discharge. The median time to readmission, if readmission occurred within 30 days, was 12 days (IQR, 5–20 days). Of all 30-day readmissions, 101 (7.9%) were due to HD access infection or complication, 136 (10.6%) were due to peritonitis and PD catheter-related infection or complication, 134 (10.5%) were due to other infection, 166 (13.0%) were due to cardiovascular causes, 81 (6.3%) were due to electrolyte and fluid disorders, and 39 (3.0%) were due to complication of a transplanted kidney. All other diagnoses made up 625 (48.8%) of 30-day readmissions.
Univariate odds ratios (ORs) and associated 95% confidence intervals (CIs) for all-cause 30-day readmission are shown in Table 2. In univariate analysis, age older than 2 years at admission (OR, 0.6; 95% CI, 0.5 to 0.8) and admissions with an operating room charge (OR, 0.7; 95% CI, 0.6 to 0.9) were associated with a lower odds of 30-day readmission. Admitted patients with three or more comorbid conditions (OR, 1.4; 95% CI, 1.0 to 2.0) and admissions with a length of stay>14 days (OR, 1.3; 95% CI, 1.0 to 1.7) were associated with a higher risk of 30-day readmission. Sex, race, dialysis modality, principal admitting diagnosis, discharge season, and an intensive care unit charge were not associated with 30-day readmission.
Results of a multivariate model for all-cause 30-day readmission are shown in Table 4. Older age at index admission (OR, 0.6; 95% CI, 0.5 to 0.8) and an associated operating room charge (OR, 0.6; 95% CI, 0.5 to 0.8) were independently associated with a lower risk of readmission within 30 days of discharge. Length of stay>14 days (OR 1.5; 95% CI, 1.1 to 2.0) and HD were associated with higher odds of readmission (OR, 1.2; 95% CI, 1.0 to 1.4).
Table 4: Multivariate mixed effects logistic regression for all-cause and similar-cause 30-day readmission
30-Day Readmission for Similar Cause
Of all 30-day readmissions, 658 (51%) were for a principal diagnosis in the same diagnosis group as the index admission. In the multivariate model, older age at admission was associated with a lower risk of readmission with the same diagnosis group (OR, 0.6; 95% CI, 0.5 to 0.9). HD (OR, 1.4; 95% CI, 1.1 to 1.7) and length of stay≥14 days (OR, 2.3; 95% CI, 1.6 to 3.4) were associated with higher odds of readmission with the same diagnosis group. None of the seven principal diagnosis groups examined were significantly associated with readmission with the same diagnosis group as the index admission 30 days following discharge (Table 4).
Discussion
We observed that 25% of hospital admissions of children with ESRD receiving chronic dialysis were followed by a readmission within 30 days of discharge. Admissions with length of stay≥14 days were more likely to be followed by a readmission. An operating room charge during the index hospitalization was associated with a lower risk of readmission. Although 50% of the readmissions were within the same diagnosis group as the index admission, a particular diagnosis group was not associated with readmission. Patients with four or more admissions during the observation period were more likely to be younger, to be black, to have public insurance, and to have more comorbid conditions compared with patients with fewer admissions. However, of these factors, only age younger than 2 years was associated with a higher risk of all-cause and similar-cause 30 day readmission. These findings are consistent with previous work that examined clinical characteristics of children recurrently admitted to children’s hospitals (2).
The high rate of readmission we observed (25%) in the PHIS hospitals is consistent with, but slightly lower than, data from the most recent US Renal Data System (USRDS) report (35% readmission rate) in patients receiving chronic dialysis (1,9,10). It is possible that the older age of the USRDS cohort lead to this difference. The 30-day readmission rate for children receiving long-term dialysis is far higher than the 6% rate reported for all hospitalized children (8). In that study, diseases of the genitourinary system were grouped together and had a 30-day unplanned readmission rate of 15%. The 30-day readmission rate observed in our study, and in the USRDS data, is higher than the 30-day readmission rate of 20% observed in children with neoplasms, another population with a high prevalence of central line usage (8). Alarmingly, the readmission rate we observed in children with ESRD is at least as high as the 30-day readmission rates for 80-year-old adults with heart failure (25%), acute myocardial infarction (20%), or pneumonia (18%) (11). This demonstrates the complexity of children with ESRD when compared with clinically fragile populations.
The higher readmission rate among the youngest children receiving dialysis is consistent with the general pediatric population, where those younger than 1 year of age had a 9% increase in 30-day readmission compared with children older than 13 years of age (8). Furthermore, a longer length of stay during the index admission has been associated with a higher risk of readmission (8). Lengths of stay≥14 days may represent higher patient acuity, higher admission complexity, or a complication during the admission, making the association of longer length of stay with a higher likelihood of readmission complex. Admissions with an operating room charge may reflect planned, primarily procedural index admission in a relatively stable patient, and the lower odds of readmission following these admissions is probably not causal.
Although the validity of using 30-day readmission as a quality and/or payment metric remains debated (12,13), readmissions of chronic dialysis patients may become a quality measure in the near future (5). It is unknown whether readmissions are a marker of poor hospital care, poor outpatient care (14), suboptimal interactions between discharging hospitals and outpatient dialysis facilities (9), or appropriate and necessary care in a medically complex population (13). However, hospitalizations increase the cost of ESRD care, create a burden on the patient and family, and increase the risk of iatrogenesis and medical complication. Hospitalizations are also a marker of more severe illness acuity. Optimizing medical care across the complex medical system to prevent hospitalizations and readmissions is a prudent future goal.
Readmissions may be lowered by targeting the conditions most commonly resulting in hospitalization, including peritonitis, cardiovascular events, and HD access infections and complications. For example, the use of atriovenous fistulas or grafts is associated with lower rates of hospitalization. Ma et al. (15) found that patients with a fistula had a hospitalization rate of 0.4%, while children dialyzed with a central catheter had an admission rate of 3%. Ng et al. also found a higher rate of hospitalizations among children with central venous catheters (16). The National Kidney Foundation recommends that pediatric patients on maintenance HD be dialyzed using permanent dialysis access (17). However, permanent access placement in children is complicated by patient size, hospital resources, and the pediatric preference for deceased-donor kidneys. Therefore, almost 80% of children initiating HD do so with a catheter (18). Achievement of quality measures in adolescent HD patients (19) and closer monitoring after hospitalization in adult HD patients were associated with lower hospitalization rates (20). Interventions that target decreasing peritonitis rates, catheter-related bloodstream infections, and the highest risk patients may also decrease hospitalizations and readmissions (10,21–25).
Cardiovascular disease is a major cause of morbidity and mortality in children with ESRD (26) and accounted for approximately 16% of all admission in our cohort. Monitoring of a hematocrit-guided ultrafiltration algorithm lowered hospitalization for fluid overload and hypertension among pediatric HD patients (27). Cardiovascular admissions are higher on the day following the longest interdialytic interval in HD patients (28,29); therefore, alternative dialysis schedules deserve further study to determine their impact on hospitalization and readmission (30,31).
We acknowledge the limitations of using administrative data for the evaluation of clinical outcomes. Several important variables, including the type of HD access and the presence of a PD catheter at time of admission, could not be readily discerned from the PHIS dataset. Dialysis vintage is also unknown, and rate of admission around the time of dialysis initiation tends to be higher (1). Indeed, patients who develop ESRD during the index hospitalization would be included in our cohort. Per our inclusion criteria, all admissions had a procedure code for HD or PD in addition to an ESRD code (Table 1). Therefore, any patient who was admitted but did not receive a dialysis treatment during the hospitalization would be missed in our analysis. Our data did not contain information on patient death outside of the hospital or transplantation; therefore, the competing risks of death and transplantation were not included in the evaluation. In contrast to prior reports in the general pediatric population, we could not identify whether readmissions were planned (8); however, >50% of the admissions seem unplanned by their principal diagnoses (e.g., hypertension, peritonitis, HD catheter infection). The PHIS database presents a unique opportunity to investigate less common diseases in children using its large number of admissions. However, this database has not been previously validated in chronic dialysis patients, and it is limited by available coding. For a large proportion of admissions (30%), data on the underlying cause of ESRD (glomerular versus congenital/obstructive) were missing, which was uninformative. We believe that any coding inaccuracies are random in nature and would not bias the results. Finally, patients readmitted to a hospital not reporting data to PHIS would be missed in our analysis. We attempted to control for this by only including patients younger than 17 years of age to reduce the chance that a patient would have transitioned out of a children’s hospital.
In conclusion, expenditures for care of long-term dialysis patients are very high, leading payers to seek methods for lowering costs. From 2002 to 2006, hospitalizations of all patients with ESRD in the United States cost $31.5 billion, which was one third of all ESRD costs (1). We observed that a significant number of admissions (25%) in pediatric long-term dialysis patients are followed by readmission within 30 days. Further research is needed to clearly differentiate preventable from unpreventable admissions and to clarify the necessity and possible benefits of hospitalization despite their added costs. Future avenues of research include prospective evaluation and intervention studies to reduce readmission in this patient population. Interventions could focus on inpatient, outpatient, or the transition of care as areas of possible intervention. Studies may benefit from targeting patients at highest risk of readmission, as demonstrated by our findings.
Disclosures
None.
Acknowledgments
B.L. is supported by a Career Development award in Comparative Effectiveness Research from the University of Pennsylvania and National Institutes of Health (1KM1CA156715-01). S.F. is supported by a grant from the National Institute of Diabetes and Digestive and Kidney Diseases (K24DK078737).
Published online ahead of print. Publication date available at www.cjasn.org.
This article contains supplemental material online at http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN.05410513/-/DCSupplemental.
See related editorial, “The Role of 30-Day Readmission as a Measure of Quality,” on pages 440–442.
References
1. USRDS: US Renal Data System, USRDS 2012 Annual Data Report: Atlas of Chronic Kidney Disease and End-Stage Renal Disease in the United States, Bethesda, MD, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 2012
2. Berry JG, Hall DE, Kuo DZ, Cohen E, Agrawal R, Feudtner C, Hall M, Kueser J, Kaplan W, Neff J: Hospital utilization and characteristics of patients experiencing recurrent readmissions within children’s hospitals. JAMA 305: 682–690, 2011
3. U.S. Department of Health and Human Services, HRSA, Maternal and Child Health Bureau: Child Health USA 2011, Rockville, Maryland, U.S. Department of Health and Human Services, 2011
4. Daratha KB Short RA, Corbett CF, Ring ME, Alicic R, Choka R, Tuttle KR: Risks of subsequent hospitalization and death in patients with kidney disease. Clin J Am Soc Nephro 3: 409–416, 2012
5. MedPAC: MedPAC letter to CMS. Available at:
http://www.medpac.gov/documents/08312012_ESRD_comments.pdf. Accessed January 22, 2013
6. Gay JC, Hain PD, Grantham JA, Saville BR: Epidemiology of 15-day readmissions to a children's hospital. Pediatrics 127: e1505–1512, 2011
7. Feudtner C, Levin JE, Srivastava R, Goodman DM, Slonim AD, Sharma V, Shah SS, Pati S, Fargason C Jr, Hall M: How well can hospital readmission be predicted in a cohort of hospitalized children? A retrospective, multicenter study. Pediatrics 123: 286–293, 2009
8. Berry JG, Toomey SL, Zaslavsky AM, Jha AK, Nakamura MM, Klein DJ, Feng JY, Shulman S, Chiang VW, Kaplan W, Hall M, Schuster MA: Pediatric readmission prevalence and variability across hospitals. JAMA 309: 372–380, 2013
9. He K, Kalbfleisch JD, Li Y, Li Y: Evaluating hospital readmission rates in dialysis facilities; adjusting for hospital effects. Lifetime Data Anal 19: 490–512, 2013
10. Wingard RL, Chan KE, Hakim R: RightReturn. Partnering to reduce the high rate of hospital readmission for dialysis-dependent patients. Nephrol News Issues 26: 20–22, 2012
11. Dharmarajan K, Hsieh AF, Lin Z, Bueno H, Ross JS, Horwitz LI, Barreto-Filho JA, Kim N, Bernheim SM, Suter LG, Drye EE, Krumholz HM: Diagnoses and timing of 30-day readmissions after hospitalization for heart failure, acute myocardial infarction, or pneumonia. JAMA 309: 355–363, 2013
12. Srivastava R, Keren R: Pediatric readmissions as a hospital quality measure. JAMA 309: 396–398, 2013
13. Vaduganathan M, Bonow RO, Gheorghiade M: Thirty-day readmissions: The clock is ticking. JAMA 309: 345–346, 2013
14. Lorch SA, Baiocchi M, Silber JH, Even-Shoshan O, Escobar GJ, Small DS: The role of outpatient facilities in explaining variations in risk-adjusted readmission rates between hospitals. Health Serv Res 45: 24–41, 2010
15. Ma A, Shroff R, Hothi D, Lopez MM, Veligratli F, Calder F, Rees L: A comparison of arteriovenous fistulas and central venous lines for long-term chronic haemodialysis. Pediatr Nephrol 28: 321–326, 2013
16. Ng LJ, Chen F, Pisoni RL, Krishnan M, Mapes D, Keen M, Bradbury BD: Hospitalization risks related to vascular access type among incident US hemodialysis patients. Nephrol Dial Transplant 26: 3659–3666, 2011
17. NKF: Clinical Practice Guidelines and Clinical Practice Recommendations. 2006. Available at:
http://www.kidney.org/professionals/KDOQI/guideline_upHD_PD_VA/va_rec8.htm. Accessed January 28, 2013
18. North American Pediatric Renal Trials and Collaborative Studies. 2011 Annual Report. Available at:
https://web.emmes.com/study/ped/annlrept/annualrept2011.pdf. Accessed 1/22/2013
19. Gorman G, Neu A, Fivush B, Frankenfield D, Furth S: Hospitalization rates and clinical performance measures in U.S. adolescent hemodialysis patients. Pediatr Nephrol 25: 2335–2341, 2010
20. Chan KELJ, Lazarus JM, Wingard RL, Hakim RM: Association between repeat hospitalization and early intervention in dialysis patients following hospital discharge. Kidney Int 76: 331–341, 2009
21. Chua AN, Goldstein SL, Bell D, Brewer ED: Topical mupirocin/sodium hypochlorite reduces peritonitis and exit-site infection rates in children. Clin J Am Soc Nephrol 4: 1939–1943, 2009
22. Corrigan ML, Pogatschnik C, Konrad D, Kirby DF: Hospital readmissions for catheter-related bloodstream infection and use of ethanol lock therapy: Comparison of patients receiving parenteral nutrition or intravenous fluids in the home vs a skilled nursing facility. JPEN J Parenter Enteral Nutr 37: 81–84, 2013
23. Furth SL, Donaldson LA, Sullivan EK, Watkins SLNorth American Pediatric Renal Transplant Cooperative Study: Peritoneal dialysis catheter infections and peritonitis in children: A report of the North American Pediatric Renal Transplant Cooperative Study. Pediatr Nephrol 15: 179–182, 2000
24. Holloway M, Mujais S, Kandert M, Warady BA: Pediatric peritoneal dialysis training: characteristics and impact on peritonitis rates. Perit Dial Int 21: 401–404, 2001
25. Lane JC, Warady BA, Feneberg R, Majkowski NL, Watson AR, Fischbach M, Kang HG, Bonzel KE, Simkova E, Stefanidis CJ, Klaus G, Alexander SR, Ekim M, Bilge I, Schaefer FInternational Pediatric Peritonitis Registry: Relapsing peritonitis in children who undergo chronic peritoneal dialysis: A prospective study of the international pediatric peritonitis registry. Clin J Am Soc Nephrol 5: 1041–1046, 2010
26. Mitsnefes MM: Cardiovascular disease in children with chronic kidney disease. J Am Soc Nephrol 23: 578–585, 2012
27. Goldstein SL, Smith CM, Currier H: Noninvasive interventions to decrease hospitalization and associated costs for pediatric patients receiving hemodialysis. J Am Soc Nephrol 14: 2127–2131, 2003
28. Foley RN, Gilbertson DT, Murray T, Collins AJ: Long interdialytic interval and mortality among patients receiving hemodialysis. N Engl J Med 365: 1099–1107, 2011
29. Springel T, Laskin B, Shults J, Keren R, Furth S: Longer interdialytic interval and cause-specific hospitalization in children receiving chronic dialysis. Nephrol Dial Transplant 28: 2628–2636, 2013
30. Chertow GM, Levin NW, Beck GJ, Depner TA, Eggers PW, Gassman JJ, Gorodetskaya I, Greene T, James S, Larive B, Lindsay RM, Mehta RL, Miller B, Ornt DB, Rajagopalan S, Rastogi A, Rocco MV, Schiller B, Sergeyeva O, Schulman G, Ting GO, Unruh ML, Star RA, Kliger ASFHN Trial Group: In-center hemodialysis six times per week versus three times per week. N Engl J Med 363: 2287–2300, 2010
31. Walsh M, Culleton B, Tonelli M, Manns B: A systematic review of the effect of nocturnal hemodialysis on blood pressure, left ventricular hypertrophy, anemia, mineral metabolism, and health-related quality of life. Kidney Int 67: 1500–1508, 2005
