Potential Impact of Nutritional Intervention on End-Stage Renal Disease Hospitalization, Death, and Treatment Costs

ORIGINAL RESEARCH Potential Impact of Nutritional Intervention on End-Stage Renal Disease Hospitalization, Death, and Treatment Costs Eduardo Lacson, Jr., MD,* T. Alp Ikizler, MD, J. Michael Lazarus, MD,* Ming Teng, MD,* and Raymond M. Hakim, MD* Objective: Our objective was to estimate the effect of an improvement in nutrition, represented by albumin concentrations, on hospitalization, mortality, and Medicare end-stage renal disease (ESRD) program cost. Design: Based on published trials, the impact of an improvement in serum albumin of 0.2 g/dl from a hypothetical nutritional program for severely malnourished patients with albumin 3.5 g/dl (base case) was estimated by reassigning patients to higher albumin categories, along with outcome risks associated with the new albumin category. Setting: Data from Fresenius Medical Care North America (Waltham, MA) were utilized in regression models to determine the association between albumin and change in albumin concentration with outcomes. Results: Albumin 3.5 g/dl was associated with a 2-fold increase in death and hospitalization risk, compared to 4 g/dl (P.001) in this population. An increase in albumin concentration was associated with a lower risk of death and hospitalization, whereas a declining albumin concentration led to worse outcomes. Projections for the United States dialysis population from the base case showed 1400 lives saved, 6000 hospitalizations averted, and $36 million in Medicare cost savings resulting from a reduction of 20,000 hospital days. A sensitivity analysis, varying the albumin response to 0.1 and 0.3 g/dl combined with varying albumin responder rates to 25% and 75% of patients, revealed robust results. Conclusion: Nutritional interventions that increase serum albumin by 0.2 g/dl (e.g., via oral nutritional supplements) may lead to considerable improvements in mortality, hospitalization, and treatment costs. 2007 by the National Kidney Foundation, Inc. NUTRITIONAL SUPPLEMENTATION, whether in the form of intradialytic parenteral nutrition (IDPN), 1-3 daily oral dietary assistance or supplementation, 4-6 or intradialytic oral nutrition (IDON), 7-9 increases the *Fresenius Medical Care North America, Waltham, Massachusetts. Nephrology Division, Department of Internal Medicine, Vanderbilt University Medical Center, Nashville, Tennessee. Address reprint requests to Eduardo Lacson, Jr., MD, MPH, FASN, Vice President of Clinical Science, Epidemiology, and Research, Fresenius Medical Care North America, 920 Winter St., Waltham, MA 02451-1457. E-mail: elacsonj@fmc-na.com 2007 by the National Kidney Foundation, Inc. 1051-2276/07/1706-0001$32.00/0 doi:10.1053/j.jrn.2007.08.009 albumin concentration in end-stage renal disease (ESRD) patients. These nutritional intervention studies indicate that serum albumin improves, on average, by 0.2 g/dl, ranging from 0.1 to 0.3 g/dl in the majority of patients. To our knowledge, a focused analysis that quantifies how graded improvements in serum albumin levels in patients with severe uremic wasting will influence hospitalization, death rates, and potential economic consequences is not available in the nephrology literature. We utilized the Fresenius Medical Care North America (Waltham, MA) database to estimate how nutritional improvement, indicated by increased serum albumin levels, may affect hospitalization, Journal of Renal Nutrition, Vol 17, No 6 (November), 2007: pp 363 371 363

364 LACSON ET AL mortality, and treatment costs for the United States (US) ESRD population. Methods The population examined includes 77,205 patients from the Fresenius Medical Care, North America (FMCNA) database with laboratory test results from October 1 to December 31, 2003 (fourth quarter, 2003) who survived into January 1, 2004. The case mix (age, gender, race, presence or absence of diabetes mellitus, and dialysis vintage), body surface area (BSA), and laboratory information (albumin, equilibrated Kt/V, hemoglobin, calcium, phosphorus, creatinine, and white blood cell count) were collected for analysis. Vascular access data were obtained for the hospitalization analyses. Demographic characteristics and laboratory profiles of the patient cohort and analytical subsets are presented as mean standard deviation or as percentages of the total. Dialysis vintage was defined as the number of days from the first maintenance dialysis on record, and was transformed into a square root in the analytical models for linearity. Follow-up information was tracked in the succeeding year, for hospitalizations, discharges (e.g., transplant, or transfer to another facility), and deaths from January 1 to December 31, 2004. The patient information was nationally distributed and comprised 26% of the US dialysis population in 2004. The analysis was conducted in three phases. Phase 1 examined associations between serum albumin and mortality risk, as well as hospitalization risk, consistent with previous studies. 10-12 The mean albumin concentration from the fourth quarter of 2003 for each of 77,205 patients was used in categorizing albumin (by increments of 0.1 g/dl) into ranges from 3.20 g/dl, 3.21 to 3.30 g/dl, and up to 4.00 g/dl. Three Cox proportional hazard models were constructed to determine mortality risk, by (1) using albumin as a single predictor variable, (2) adding case-mix (CM) variables to the albumin model, and then (3) adding laboratory (Lab) variables to model 2. The BSA was included in the final model under the CM Lab adjustment. For hospitalization, three logistic regression models were similarly constructed, except that creatinine and calcium were replaced by vascular access (VA) type as a predictor variable in the final model. Phase 2 utilized change in albumin concentration over an extended 1-year baseline period (2003) as an independent variable to determine an association with risk of hospitalization and death in 2004 for the subset of patients in the original Phase 1 cohort who were already on dialysis as of the first quarter of 2003, and who had a mean serum albumin of 3.5 g/dl. The change was defined as the difference from the mean serum albumin in the fourth quarter minus the mean serum albumin in the first quarter. Three models were created similar to that in Phase 1, with unadjusted, CM-adjusted, and CM Lab adjusted models to illustrate the associated mortality and hospitalization risk profiles. Phase 3 utilized a theoretical model on the basis of a hypothetical systematic nutritional intervention that increases serum albumin concentration for selected patients with severe malnutrition, defined (based on K/DOQI) as mean serum albumin concentration of 3.5 g/dl, 13 during the fourth quarter of 2003. Patients who respond are transferred from one albumin category to a higher albumin category, and the model assumes that the mortality and hospitalization risk for the higher category becomes applicable. We utilized crude (unadjusted) rates for both the hospitalization and mortality obtained from Phase 1 for each albumin category, at increments of 0.1 g/dl. In the base case, the projected increase of albumin concentration was an average of 0.2 g/dl for at least 50% of the intervention population. Therefore, 50% of the intervention patients from each albumin category ( 3.5 g/dl) were upgraded by 0.2 g/dl (e.g., from 3.41-3.50 g/dl to 3.61-3.70 g/dl, and from 3.31-3.40 g/dl to 3.51-3.60 g/dl) as depicted in Figure 1. The death risk associated with each albumin category was then applied to the projected (redistributed) patient population. A similar projection was utilized for hospitalization. The overall difference in death and hospitalization rates between the actual and projected number of patients in each serum albumin category represented the potential impact of the hypothetical intervention. The estimates were then extrapolated to the prevalent US dialysis population, assuming that the FMCNA population comprised a proportionately distributed national sample. The sensitivity analyses consisted of varying key assumptions along two dimensions: (1) net impact of the intervention on serum albumin, from 0.1 g/dl to 0.3

NUTRITIONAL INTERVENTION AND OUTCOMES IN ESRD 365 Figure 1. Projected distribution shift of the patient population resulting from a hypothetical systematic intervention with a 0.2 g/dl improvement in serum albumin in 50% of patients (base case). Only patients in albumin categories of 3.50 g/dl were projected to be reassigned to the higher albumin category, whereas the two categories between 3.51 and 3.70 g/dl were projected to receive some of those patients who improved. g/dl, and (2) change in prevalence of the improvement of albumin in patients with a baseline 3.5 g/dl to either 25% or 75%. Projected outcomes included mortality rate, lives saved, hospitalization events, and hospital days. For US hospitalization cost estimates, the projected counts of avoided hospital days were multiplied by the average Medicare cost of $1800 per hospital day. 14 The method used to derive this Medicare cost is in the Appendix. Statistical analyses were performed using SAS version 9.1 (SAS, Inc., Cary, NC). Results The patient characteristics for the mortality (N 77,205) and hospitalization (N 67,627) cohort populations in Phase 1 and Phase 2 (patient subset of Phase 1) are shown in Table 1. The mortality and hospitalization cohorts in each phase of the study were virtually identical in all aspects considered. In Phase 2 of the study, patients had a mean serum albumin of 3.22 0.30 g/dl (first quarter), and the surviving cohort improved the population mean albumin to 3.5 0.4 g/dl by the fourth quarter of 2003. By definition, these patients had to have survived from the first quarter into the fourth quarter of 2003, and thus patients who had serum albumins 3.5 g/dl from the first quarter of 2003 who failed to survive were excluded: hence the unavoidable selection bias toward patients whose albumins improved. However, the direction of such a bias in relation to outcomes would have been toward the null, because patients who died were more likely to have had lower albumin concentrations. Compared with the overall cohorts from Phase 1, the subset of patients in Phase 2 were older, more likely to be female, predominantly of white race, and more likely to be diabetic, and had a smaller BSA, less vintage, and a greater prevalence of patients with hemodialysis catheter access. The albumin-associated risk profiles from Phase 1 are shown in Figure 2. Albumin exhibited a robust association with both mortality and hospitalization risk. Albumin consistently maintained the strongest association with either outcome when compared with all other predictor variables included in any model (data not shown). Compared with patients with albumin 4.0 g/dl, patients with a mean serum albumin 3.5 g/dl exhibited an increased unadjusted risk for both mortality and hospitalization of approximately 3- to 5-fold (P.001). These risks were attenuated with CM Lab adjustment by an order of magnitude, down to 2- to 4-fold for both mortality and hospitalization (P.001). Phase 2 results are depicted in Figure 3, and confirm the association between a change in albumin concentration and risk of mortality and hospitalization. A decline in serum albumin concentration of 0.10 to 0.19 g/dl for severely malnourished patients increased the associated CM Lab-adjusted mortality risk by 9% (P.001), and by as much as 88% for a decline of 0.3 g/dl (P.001), as shown in Figure 3a. An increase in mean serum albumin concentration by 0.1 g/dl was associated with a significant 16% improved survival in the CM Lab-adjusted model (P.05), and by as much as a 30% improved survival with a 0.3 g/dl increase (P.001). In the hospitalization cohort, a decline of 0.3 g/dl had an associated increase in unadjusted hospitalization risk of 104% (P.001), attenuated in the fully adjusted models to 41% (P.05), as shown in Figure 3b. However, when albumin was increased by 0.2 g/dl, the associated hospitalization risk decreased by 25% for both unadjusted and fully adjusted models (range, 24% to 28%; P.05). Further, in patients whose albumin increased by

366 LACSON ET AL Table 1. Patient Characteristics for Mortality and Hospitalization Phase 1 (All Patients) and Phase 2 (Patient Subset With Mean Albumin 3.5 g/dl During the First Quarter of 2003) Patient Characteristics* Phase 1 Phase 2 Mortality Hospitalization Mortality Hospitalization Patient counts 77,205 67,627 8844 7783 Age (years) 61.4 (15.0) 61.6 (14.9) 63.3 (14.9) 63.3 (14.4) % female 47.0% 47.6% 54.1% 54.9% Race % white 49.2% 49.0% 52.2% 51.8% % black 40.8% 41.1% 38.1% 38.4% % other 10.0% 9.9% 9.7% 9.8% % with diabetes 51.9% 52.7% 63.8% 64.1% Vintage (days) 1254 1269 1110 1115 Body surface area (m 2 ) 1.83 (0.27) 1.83 (0.27) 1.78 (0.26) 1.79 (0.26) Access type % with fistula 36% 36% 30% 30% % with graft 39% 40% 41% 41% % with catheter 23% 23% 28% 28% Albumin (g/dl) 3.87 (0.42) 3.87 (0.41) 3.54 (0.41) 3.55 (0.40) Equilibrated Kt/V 1.42 (0.29) 1.43 (0.28) 1.43 (0.29) 1.44 (0.28) Hemoglobin (g/dl) 11.8 (1.2) 11.8 (1.2) 11.8 (1.2) 11.8 (1.1) Calcium (mg/dl) 9.0 (0.8) 9.0 (0.8) 8.8 (0.7) 8.8 (0.7) Phosphorus (mg/dl) 5.3 (1.4) 5.3 (1.4) 5.3 (1.5) 5.3 (1.5) Creatinine (mg/dl) 9.5 (3.4) 9.5 (3.3) 8.4 (2.8) 8.4 (2.8) WBC count 7.2 (2.4) 7.2 (2.4) 7.3 (2.4) 7.3 (2.4) *Presented as mean (standard deviation) or percentage of the total where indicated, as of the end of the fourth quarter of 2003. Access type hemodialysis vascular access type, with a total of 100% because of missing data. WBC count white blood cell count, in units of 10 3 per microliter. 0.4 g/dl, the hospitalization risk declined by 41% (P.001). To our knowledge, there was no operative system-wide intervention to improve serum albumin in this subset of patients, except for general dietary recommendations. Among the 77,205 patients, there were 12,929 (17%) with a mean serum albumin 3.5 g/dl in the fourth quarter of 2003, meeting the definition of severe malnutrition, i.e., the target population for the hypothetical nutritional intervention. In the base case, if 50% of the target patients could improve serum albumin by 0.2 g/dl, the predicted overall death rate would decline by (an absolute) 0.48%, corresponding to 369 deaths averted. Similar base-case projections were made from the hospitalization information, to obtain an estimated possible 1646 hospitalizations averted and 5247 hospital days saved. The results are summarized in Table 2, with a sensitivity analysis on achieving the 0.2 g/dl improvement in serum albumin in only 25% as well as in 75% of target patients. The Phase 3 extrapolations of FMCNA projections for the US dialysis population are shown in Tables 3 to 6. A focused intervention that could improve albumin for all dialysis patients in the US with severe malnutrition could save 1400 lives per year, considering only the basecase scenario, i.e., 50% of severely malnourished patients improving their serum albumin by 0.2 g/dl. The sensitivity analysis in Table 3 shows that the projected number of lives saved, based on the achievable albumin improvement of 0.2 g/dl, is 712 lives for a modest 25% effect, 1419 lives for a likely 50% effect, and 2127 patients for a plausible 75% effect. Similarly, the projected number of possible lives saved for a 50% response to the intervention if the improvement in albumin is a modest 0.1 g/dl is 827 lives, while for 0.3 g/dl this increases to 1869 lives (with other combinations also shown in Table 3). The projected impact of the base case on hospitalization is a possible reduction by 6331 hospitalization events and 20,181 fewer hospital days. The sensitivity analyses are provided in Tables 4 and 5. Finally, the reduction in hospitalization was estimated to lead to a potential savings of $36 million in the base case, with a range from $14 million if only 25% of patients responded with an increase in serum albumin of 0.1 g/dl, to as high as $88 million if 75% of patients responded

NUTRITIONAL INTERVENTION AND OUTCOMES IN ESRD 367 Figure 2. Association between mean serum albumin level in the fourth quarter of 2003 with (a) mortality risk and (b) hospitalization risk during 2004, each shown with results from unadjusted, case-mix (CM)-adjusted, and fully adjusted models. with an improvement in serum albumin of 0.3 g/dl, as depicted in Table 6. Discussion This study confirms the strong association between albumin and change in albumin concentrations with outcomes. We further provide a model whereby small, realistic, but clinically feasible improvements in serum albumin were speculatively translated into significant numbers of lives saved, averted hospitalizations, and substantial cost savings. The current model allows for a graded change in albumin levels as well as the corresponding risks, without simply relying on all patients reaching a single threshold (e.g., all patients albumin 3.8 g/dl). 15 Furthermore, the model assumes a plausible response to the intervention, ranging from 25% to 75% of target patients with the base case at a 50% response, consistent with the low-end predicted response in another study. 16 The magnitude of albumin increase in the base case was 0.2 g/dl, achievable within the realm of published experience from a variety of nutritional interventions in ESRD. 1-9 Among enteral interventions, the effect on serum albumin was an average of 0.23 g/dl in a recent meta-analysis of 18 studies. 6 Our model indicates that an improvement in albumin of as little as 0.1 g/dl achieved in only 25% of severely malnourished (baseline serum albumin, 3.5 g/dl) ESRD patients in the US could still lead to 415 lives saved, 2165 hospitalizations averted, and approximately $15 million in potential annual cost savings, which is particularly strik-

368 LACSON ET AL Figure 3. Change in mean serum albumin level (first vs. fourth quarter of 2003) in the patient subset that started the first quarter of 2003 at a mean serum albumin of 3.5 g/dl, and the association with (a) mortality risk and (b) hospitalization risk during 2004, each shown with results from unadjusted, case-mix (CM)- adjusted, and fully adjusted models. Table 2. Projections of Possible Impact of a Systematic Intervention That Improves Albumin by 0.2 g/dl in 25% to 75% of Patients With a Baseline Serum Albumin 3.5 g/dl From the FMCNA Dialysis Population FMCNA Population, Possible Impact of Intervention % of Patients With Albumin 3.5 Improved by 0.2 g/dl 25% 50% 75% Number of hospitalization potential events avoided (%) 823 (0.80%) 1646 (1.59%) 2468 (2.59%) Potential hospital days avoided (%) 2624 (0.67%) 5247 (1.35%) 7871 (2.02%) Potential percentage decline in crude death rate 0.24% 0.48% 0.72% Potential lives saved 189 369 553 FMCNA, Fresenius Medical Care, North America.

NUTRITIONAL INTERVENTION AND OUTCOMES IN ESRD 369 Table 3. Projections for the US ESRD Population Regarding Possible Lives Saved With an Intervention That Improves Nutritional Status in Patients With Serum Albumin 3.5 g/dl Possible Lives Saved Penetrance of Improvement in Albumin Improved Albumin From 3.5 g/dl 25% 50% 75% 0.1 g/dl 415 827 1,242 0.2 g/dl 712 1,419 2,127 0.3 g/dl 935 1,869 2,800 US ESRD, United States end-stage renal disease. Table 4. Projections for the US ESRD Population Regarding Possible Averted Hospitalizations With an Intervention That Improves Nutritional Status in Patients With Serum Albumin 3.5 g/dl Possible Averted Hospitalizations Penetrance of Improved Albumin Improvement in Albumin From 3.5 g/dl 25% 50% 75% 0.1 g/dl 2,165 4,831 6,496 0.2 g/dl 3,165 6,331 9,492 0.3 g/dl 4,815 9,865 14,450 US ESRD, United States end-stage renal disease. Table 5. Projections for the US ESRD Population Regarding Possible Reduction in Hospital Days With an Intervention That Improves Nutritional Status in Patients With Serum Albumin 3.5 g/dl Possible Reduction in Hospital Days Penetrance of Improved Albumin Improvement in Albumin from 3.5 g/dl 25% 50% 75% 0.1 g/dl 8,038 16,077 24,112 0.2 g/dl 10,092 20,181 30,273 0.3 g/dl 16,238 32,481 48,719 US ESRD, United States end-stage renal disease. ing considering that such an improvement is clearly achievable with available nutritional interventions. The most optimistic scenario, incorporating a mean increase of 0.3 g/dl in 75% of patients, projects potentially 2800 lives saved, 14,450 hospitalizations averted, and $88 million in potential annual cost savings. Optimal nutritional support in ESRD patients is not well-defined, and can be provided on several levels. Patient-related barriers include poor nutritional knowledge, poor appetite, lack of assistance with shopping and cooking, low fluid intake, acidosis, inadequate dialysis, depression, gastrointestinal symptoms, and difficulty chewing or swallowing. 17 In addition, the catabolic effect of comorbid conditions (e.g., infections) and of HD, postdialysis symptoms, missed opportunities for food intake as a result of preparations for dialysis, travel busy-ness as well as travel time to the dialysis unit, and the HD procedure itself, all combine to further exacerbate the imbalance between poor dietary intake and increased metabolic requirements. 18-21 Ease of monitoring makes interventions performed within HD units attractive, although such interventions are often limited to renal dieticians providing patients with nutritional information and dietary recommendations, with limited effectiveness. 4 Daily supplemental feeding is ideal and preferable, although patient compliance is difficult to achieve. Therefore, regular nutritional supplements administered within the dialysis unit, whether in the form of IDPN or IDON, provide an excellent supervised method to maximize patient compliance and monitor therapy. The high cost of IDPN therapy, a lack of adequately powered and designed clinical trials, and the regulatory concerns remain the greatest barriers to advocating for the utilization of this potentially beneficial treatment. Providing nutritional supplements during dialysis (i.e., IDON) is attractive, especially because feeding by mouth is more physiologic (described further below) and is an effective anabolic intervention in ESRD patients with functional gastrointestinal tracts. 6 Physiologic evidence supporting IDON includes findings that dietary intake during dialysis days is much lower when compared with nondialysis days. 21 Furthermore, evidence shows that the consumption of a protein- and energy-enriched meal during HD results in a positive protein balance, to the same extent as on a nondialysis day. 22 In addition, IDON induces a protein anabolic response comparable to IDPN in the acute setting if the protein and calorie content is similar, and promotes persistent anabolic benefits for muscle protein metabolism in the post-hd phase. 9 For example, a can each of Oxepa and Nepro taken together orally during each treatment in 20 malnourished patients (mean SD albumin, 3.44 0.34 g/dl) increased albumin to 3.68 0.34 g/dl (P.001) after 4 weeks. 23 Long-term, IDON over a 6-month period improved serum albumin (3.22 0.32 g/dl to 3.65 0.26 g/dl) and other nutritional parameters (e.g., serum prealbumin concentrations and subjective global assessment) in 85 malnourished patients. 8

370 LACSON ET AL Table 6. Projections for the US ESRD Population Regarding Potential Cost Savings From an Intervention That Improves Nutritional Status in Patients With Serum Albumin 3.5 g/dl Potential Cost Savings From Reduction of Hospitalization Penetrance of Improved Albumin Improvement in Albumin From 3.5 g/dl 25% 50% 75% 0.1 g/dl $14,469,231 $28,938,462 $43,400,769 0.2 g/dl $18,166,154 $36,325,385 $54,491,538 0.3 g/dl $29,229,231 $58,465,385 $87,694,615 US ESRD, United States end-stage renal disease. The present results should be interpreted with several caveats. First, serum albumin may be confounded by inflammation, fluid status, and a slow response to nutritional interventions. 13,24 However, albumin is an excellent marker of overall health status, correlates with other nutritional markers, responds to dietary intake, and is the most significant predictor of death and hospitalization risk. 10-13,25-27 In at least one recent randomized clinical trial, serum albumin was responsive to nutritional interventions irrespective of inflammatory markers. 17 A second limitation of this study is that the shift in serum albumin level from one category to another does not necessarily confer the improved risk profile for that particular category to a specific patient. Such causation can only be determined by a prospective, randomized clinical trial, e.g., despite numerous observational studies indicating improved outcomes with increasing dialysis dose beyond single pool Kt/V 1.2, the Hemodialysis Study was unable to demonstrate such a benefit. 28 In the absence of a definitive intervention study, we based our projections on available data, using epidemiologic tools. The analysis was conservative in projecting the change in risk by narrowing albumin categories while avoiding a blanket improvement in risk based on a single albumin threshold and excluding an unrealistic 100% patient response to the intervention. Thus, the methodology and results are conservative, believable, and plausible. Finally, the projections provided here are limited to patients with serum albumin 3.5 g/dl. It would be ideal to provide nutritional interventions for all patients who do not meet the target for serum albumin levels of 4.0 g/dl. However, the prevalence of patients with albumin 4.0 g/dl in the US is very high, considering that among patients entering the ESRD program in 2004 and 2005, the mean serum albumin concentration was 3.1 g/dl. 14 Therefore, a focused intervention for patients with severe malnutrition would be more affordable. In addition, the greatest risks, and potentially the greatest effect, should be associated with interventions directed toward patients with severe malnutrition, as noted in a prospective, randomized study where the intervention consisted of daily amino-acid supplements, and subjects with severe malnutrition ( 3.5 g/dl) had a greater improvement than those in the higher albumin strata of 3.5 to 3.8 g/dl (P.01). 7 In conclusion, we present projections based on data from our large database and a hypothetical focused nutritional intervention program. The analysis indicates that any nutritional intervention that can improve albumin by 0.2 g/dl (or higher) for patients with severe malnutrition may lead to improved outcomes, and appears to be cost-effective. Oral nutritional supplementation, at least during each hemodialysis procedure, is an effective and practical approach. A Medicare allowance for the provision of IDON may provide for an improvement in albumin levels that has been lacking since albumin was first tracked as a clinical performance measure. Alternatively, easing regulatory requirements on gifts for Medicare beneficiaries may make it possible for dialysis providers to fund a nutritional intervention program. A projected improvement in nutritional state, represented by biomarkers such as serum albumin, provides a reasonable likelihood for interventions to reduce death rates, hospitalization events, and ESRD-related costs. References 1. Chertow GM, Ling J, Lew NL, et al: The association of intradialytic parenteral nutrition administration with survival in hemodialysis patients. Am J Kidney Dis 24:912-920, 1994. 2. Capelli JP, Kushner H, Camiscioli TC, et al: Effect of intradialytic parenteral nutrition on mortality rates in end-stage renal disease care. Am J Kidney Dis 23:808-816, 1994. 3. Wong P, Smith P, Rodger D: The use of intradialytic parenteral nutrition to treat malnutrition: a case study. CANNT J 13:31-46, 2003. 4. Wilson B, Fernandez-Madrid A, Hayes A, et al: Comparison of the effects of two early intervention strategies on the

NUTRITIONAL INTERVENTION AND OUTCOMES IN ESRD 371 health outcomes of malnourished hemodialysis patients. J Ren Nutr 11:166-171, 2001. 5. Hiroshige K, Sonta T, Suda T, et al: Oral supplementation of branched-chain amino acid improves nutritional status in elderly patients on chronic haemodialysis. Nephrol Dial Transplant 16:1856-1862, 2001. 6. Stratton RJ, Bircher G, Fouque D, et al: Multinutrient oral supplements and tube feeding in maintenance dialysis: a systematic review and meta-analysis. Am J Kidney Dis 46:387-405, 2005. 7. Eustace JA, Coresh J, Kutchey C, et al: Randomized double-blind trial of oral essential amino acids for dialysisassociated hypoalbuminemia. Kidney Int 57:2527-2538, 2000. 8. Caglar K, Fedje L, Dimmitt R, et al: Therapeutic effects of oral nutritional supplementation during hemodialysis. Kidney Int 62:1054-1059, 2002. 9. Pupim LB, Majchrzak KM, Flakoll PJ, et al: Intradialytic oral nutrition improves protein homeostasis in chronic hemodialysis patients with deranged nutritional status. J Am Soc Nephrol 17:3149-3157, 2006. 10. Lowrie EG, Lew NL: Death risk in hemodialysis patients: the predictive value of commonly measured variables and an evaluation of death rate differences between facilities. Am J Kidney Dis 15:458-482, 1990. 11. Owen WF Jr, Lew NL, Liu Y, et al: The urea reduction ratio and serum albumin concentration as predictors of mortality in patients undergoing hemodialysis. N Engl J Med 329:1001-1006, 1993. 12. Leavey SF, Strawderman RL, Jones CA, et al: Simple nutritional indicators as independent predictors of mortality in hemodialysis patients. Am J Kidney Dis 31:997-1006, 1998. 13. National Kidney Foundation: Kidney Disease Outcomes Quality Initiative: Clinical practice guidelines for nutrition in chronic renal failure. Am J Kidney Dis 35(suppl): S1-S140, 2000. 14. United States Renal Data System: Excerpts from the USRDS 2006 annual data report. Am J Kidney Dis 49(suppl): S1-S296, 2007. 15. Kalantar-Zadeh K, Kilpatrick RD, Kuwae N, et al: Revisiting mortality predictability of serum albumin in the dialysis population: time dependency, longitudinal changes and populationattributable fraction. Nephrol Dial Transplant 20:1880-1888, 2005. 16. Port FK, Pisoni RL, Bragg-Gresham JL, et al: DOPPS estimates of patient life years attributable to modifiable hemodialysis practices in the United States. Blood Purif 22:175-180, 2004. 17. Leon JB, Albert JM, Gilchrist G, et al: Improving albumin levels among hemodialysis patients: a community-based randomized controlled trial. Am J Kidney Dis 48:28-36, 2006. 18. Lindholm B, Wang T, Heimburger O, et al: Influence of different treatments and schedules on the factors conditioning the nutritional status in dialysis patients. Nephrol Dial Transplant 13(suppl 6):66-73, 1998. 19. Ikizler TA, Pupim LB, Brouillette JR, et al: Hemodialysis stimulates muscle and whole body protein loss and alters substrate oxidation. Am J Physiol Endocrinol Metab 282:E107-E116, 2002. 20. Kaysen GA, Kumar V: Inflammation in ESRD: causes and potential consequences. J Ren Nutr 13:158-160, 2003. 21. Burrowes JD, Larive B, Cockram DB, et al: Effects of dietary intake, appetite, and eating habits on dialysis and non-dialysis treatment days in hemodialysis patients: cross-sectional results from the HEMO Study. J Ren Nutr 13:191-198, 2003. 22. Veeneman JM, Kingma HA, Boer TS, et al: Protein intake during hemodialysis maintains a positive whole body protein balance in chronic hemodialysis patients. Am J Physiol Endocrinol Metab 284:E954-E965, 2003. 23. Kalantar-Zadeh K, Braglia A, Chow J, et al: An antiinflammatory and antioxidant nutritional supplement for hypoalbuminemic hemodialysis patients: a pilot/feasibility study. J Ren Nutr 15:318-331, 2005. 24. Kaysen GA, Chertow GM, Adhikarla R, et al: Inflammation and dietary protein intake exert competing effects on serum albumin and creatinine in hemodialysis patients. Kidney Int 60:333-340, 2001. 25. Blumenkrantz MJ, Kopple JD, Gutman RA, et al: Methods for assessing nutritional status of patients with renal failure. Am J Clin Nutr 33:1567-1585, 1980. 26. Rocco MV, Soucie JM, Reboussin DM, et al: Risk factors for hospital utilization in chronic dialysis patients. Southeastern Kidney Council (Network 6). J Am Soc Nephrol 7:889-896, 1996. 27. Pupim LB, Evanson JA, Hakim RM, et al: The extent of uremic malnutrition at the time of initiation of maintenance hemodialysis is associated with subsequent hospitalization. J Ren Nutr 13:259-266, 2003. 28. Eknoyan G, Beck GJ, Cheung AK, et al: Effect of dialysis dose and membrane flux in maintenance hemodialysis. N Engl J Med 347:2010-2019, 2002. Appendix Computation of Average Medicare Cost for Each Hospital Day in Dialysis Patients (Source: United States Renal Data Systems [USRDS] 2006 Annual Data Report) Numerator: Inpatient Hospitalization Cost (Source: USRDS Table K.2) Total Inpatient Cost Transplant Cost Transplant Pass Through $6,699,571,501 $215,700,893 $4,964,072 Therefore: Inpatient Hospitalization Cost $6,478,906,534 Denominator: Total Hospital Days (Sources: USRDS Table G.6 and Subtable G.6.5) Total Hospital Days/Patient Year Total Years at Risk 14.1 Hospital Days/Patient-Year 249,176 Total Patient Years Therefore: Total Hospital Days 3,513,382 Hospital Days Computation: Medicare Average Cost per Hospital Day Inpatient Hospitalization Cost Total Hospital Days $6,478,906,534 3,513,382 Hospital Days Therefore: Medicare Average Cost per Hospital Day $1,844