ORIGINAL RESEARCH
Medical Nutrition Therapy for Chronic Kidney Disease Improves Biomarkers and Slows Time to Dialysis Desiree de Waal, MS, RD, CD, FAND,* Emily Heaslip, MS, RD, CD,† and Peter Callas, PhD‡ Objective: To investigate whether medical nutrition therapy (MNT) provided by a registered dietitian experienced in chronic kidney disease (CKD) slows the progression of disease and improves nutrition-related biomarkers. Design: Retrospective cohort study. Subjects: The cohort included 265 participants from a regional nephrology center in a rural state; 147 of who received MNT were compared to a group that did not receive MNT and had started dialysis over a 10-year period. Intervention: MNT by a registered dietitian with expertise in CKD. Main Outcome Measure: Average time to dialysis, based on stage of CKD at baseline, was compared between groups. In addition, the effect of MNT on the change in biochemical measures for estimated glomerular filtration rate, blood urea nitrogen, albumin, CKD Mineral and Bone Disorder (MBD) markers (phosphorous, calcium, and intact parathyroid hormone) at baseline and at follow-up (dialysis initiation or most recent laboratories if dialysis was not started) was assessed. Results: MNT group had less of a decline in estimated glomerular filtration rate than the non-MNT group (0.3 vs. 9.9 mL/minute/1.73 m2, respectively) a mean difference of 9.6 (P , 0.001). When adjusted for stage using linear regression, the mean difference was greater (11.4, P , .001). Using survival analysis and Cox proportional hazards regression, the non-MNT group was 3.15 more likely to initiate dialysis. Stratified by Stages 3 and 4 that hazard ratio increased (3.47 and 3.45, respectively). Albumin and markers of CKD-MBD were more likely to be within normal limits in the MNT group. The results indicate that better outcomes occur when MNT is given at CKD Stage 3 or 4 rather than CKD Stage 5. Conclusions: Results suggest that people with CKD who received MNT were less likely to start dialysis and had improved nutritional biomarkers than participants who did not receive MNT. Ó 2016 by the National Kidney Foundation, Inc. All rights reserved.
This article has an online CPE activity available at http://education.kidney.org/content/journal-renal-nutrition-cpe-program
Introduction
M
EDICAL NUTRITION THERAPY (MNT) provided by a registered dietitian (RD) for patients with chronic kidney disease (CKD) for estimated glomerular filtration rate (eGFRs) between 15 and 50 mL/ minute/1.73 m2 has been supported by Medicare since 20021 and many health insurance companies. To improve outcomes in people with CKD, 2 reports, the Kidney Disease Outcomes Quality Initiative2-4 and Kidney Disease Improving Global Outcomes,4,5 recommend expert *
Department of Nephrology, University of Vermont Medical Center, Burlington, Vermont. † The Community Health Centers of Burlington, Burlington, Vermont. ‡ Department of Medical Biostatistics, University of Vermont, Burlington, Vermont. Financial Disclosure: The authors declare that they have no relevant financial interests. A poster of preliminary analysis of data was previously presented at the Vermont Academy of Nutrition and Dietetics 2014 Annual Meeting and at the Academy of Nutrition and Dietetics 2014 Food and Nutrition Conference and Exposition. Address correspondence to Desiree de Waal, MS, RD, CD, FAND, University of Vermont Medical Center, Nephrology, 1 South Prospect Street, Burlington, Vermont 05401. E-mail:
[email protected] Ó 2016 by the National Kidney Foundation, Inc. All rights reserved. 1051-2276/$36.00 http://dx.doi.org/10.1053/j.jrn.2015.08.002
Journal of Renal Nutrition, Vol 26, No 1 (January), 2016: pp 1-9
dietary advice for management of progression and complications of CKD. In a retrospective cohort analysis, patients had lower mortality during the first year on dialysis therapy if they had seen a dietitian more than 12 months before dialysis initiation.6 The Academy of Nutrition and Dietetics (The Academy) Evidence Analysis Library found good evidence in their 2010 Evidence Summary to support the statement that MNT provided by an RD for adult patients with CKD is effective.7 According to 2014 Centers for Medicare and Medicaid Services report, there were over 511,900 enrollees for endstage renal disease (ESRD) during the calendar year 2013.8 The United States Renal Data System, which collects and analyzes information about CKD and ESRD, demonstrated in their 2014 Annual Data Report a steady rise in ESRD Medicare expenditures since 1991.9 The annual mean expenditure for treating a patient with renal replacement therapies (dialysis and transplant) in 2009 was approximately $70,000 per patient.10 Estimated lifetime incidences of CKD are expected to continue to grow based on current treatment patterns and risk factors.11 It is, therefore, prudent to recommend services that delay the progression of CKD to ESRD. The Academy and the National Kidney Foundation Council on Renal Nutrition developed Standards of 1
2
DE WAAL ET AL
Practice and Standards of Professional Performance for Registered Dietitians in Nephrology12,13 to ensure core standards and minimum competencies for optimal nephrology care. These standards are updated to incorporate any changes in practice at regular intervals. MNT for patients with CKD addresses nutrition issues such as, but not limited to, protein energy wasting, electrolyte imbalances, mineral and bone disorders (CKD-MBD), fluid imbalance, and vitamin deficiencies. Caring for patients with CKD requires specialized knowledge to meet the nutritional challenges patients with this disease present. Improved management of CKD can delay time to dialysis, which may result in increased quality of life (QoL) and reduced health-care costs. In this study, the researchers hypothesized that those patients with CKD who received MNT from an RD with renal expertise will have improved nutritional laboratory biomarkers at follow-up and will have a longer time period to dialysis compared with those patients with CKD who did not receive MNT before dialysis start.
Methods Study Population This study was a retrospective cohort analysis of data of patients seen for MNT at a University of Vermont Medical Center (UVMMC) Nephrology Department Outpatient Clinic, and new dialysis starts at UVMMC dialysis units throughout Vermont between April 2003 and April 2013. Patients were divided into 2 groups, those that received MNT from an RD and those that did not. This study was reviewed and approved by the University of Vermont Institutional Review Board. Data Collection Baseline and follow-up data were obtained from medical records. Two patient lists were cross referenced (all new dialysis starts and all patients seen for CKD MNT) between April 2003 and April 2013. Exclusion criteria for screening patients included patients less than 18 years old, those with acute kidney injury, a kidney transplant, and those lacking sufficient laboratory data or lost to follow-up. Pre-dialysis RD care, demographic data, and laboratory data were collected using electronic medical records. The same renal RD from the nephrology clinic saw all patients who received MNT. For both study groups, baseline was considered the date of CKD diagnosis as diagnosed by a nephrologist and documented in the electronic medical records. At baseline, gender, age, race, and whether the person had diabetes were assessed. Available laboratory values of blood urea nitrogen (BUN), eGFR, albumin, calcium, phosphorous, and intact parathyroid hormone (iPTH) were collected at baseline and follow-up. Follow-up was considered as the date of dialysis initiation or the date of most recent labora-
tories if dialysis was not initiated. Laboratory values of normal were based on the UVMMC laboratory normal values.
Statistical Analysis A comparison of the 2 groups at baseline and follow-up was completed using Fisher’s exact tests for categorical variables and 2-sample t tests for continuous variables. The percentage of participants within normal ranges for BUN, calcium, phosphorous, iPTH, and albumin was compared. Analyses were also conducted stratified by stage of CKD at baseline and adjusted for stage using logistic regression for categorical variables and linear regression for eGFR. Time to dialysis was assessed using Kaplan-Meier survival estimates. Participants who did not start dialysis during the duration of the study were censored at the date of the most recent laboratory data. Cox proportional regression was used for adjusted analysis of time to dialysis. In addition to comparisons of all patients, analyses were conducted and restricted to patients with Stage 3 to 5 CKD who started dialysis more than 1 year after date of MNTor date of diagnosis. Statistical significance was defined as a 2-sided P value , .05. No adjustments were made for multiple comparisons, as recommended when a global null hypothesis is not of interest.14 Analyses were conducted using SAS v9.4 (SAS Institute Inc., Cary, North Carolina).
Results Of the 568 patients screened who started dialysis during this time period or received MNT from an RD, 303 patients were excluded (Table 1) per exclusion criteria and not used in the analysis. Most of the exclusions (63%) were those lacking laboratory data. Of the remaining 265 patients, 147 received MNT from a renal RD and 118 did not. The 2 comparison groups, those that received MNT and those that did not, were overall very similar. There were no significant differences between those who received MNT and those who did not in terms of gender, age, race, eGFR, and presence of diabetes (Table 2). To address changes in care patterns that may have occurred over time, we divided each group that came into the study by 2.5-year quarters of entry. There was a statistical significance between the groups, with higher percentage of MNT patients from the first quarter relative to the nonMNT patients (23% of MNT patients were from the first quarter, whereas only 4% of the non-MNT patients were from the first quarter). Relatively more non-MNT patients were from the third quarter (38% compared with 22% of the MNT patients). The mean eGFR (Table 3) in the MNT group was higher than the non-MNT group, but with patients in Stages 3 and 4 CKD only, the difference between the groups diminishes at baseline. The change in eGFR15 from baseline to follow-up (Table 3) showed that the MNT group had decline in eGFR of 0.3 mL/minute/1.73 m2 as compared to the
3
MNT SLOWS PROGRESSION OF CKD, IMPROVES BIOMARKERS Table 1. Summary of Criteria for Exclusion in the Analysis Exclusion Criteria
Acute kidney injury Age , 18 Transplant Lacking laboratory data Lost to follow-up Other (kidney stones, gout, GFR . 60)
All (n 5 303)
MNT (n 5 96)
Non-MNT (n 5 207)
n (%) 18 (6) 9 (3) 39 (13) 190 (63) 40 (13) 7 (2)
n (%) 0 2 (2) 25 (26) 34 (54) 25 (26) 7 (7)
n (%) 18 (9) 7 (3) 14 (7) 156 (75) 15 (7) 0
GFR, glomerular filtration rate; MNT, medical nutrition therapy.
non-MNT group with a decline of 9.9 mL/minute/ 1.73 m2, giving a significant mean difference in decrease of 9.6 (2-sample t test P , .001). The mean difference in decline of eGFR between groups was greater when the comparisons were adjusted for stage using linear regression (11.4, P ,.001). Patients in Stages 3 and 4 CKD only, the decline in eGFR from baseline was 1 in MNT group versus 13.5 mL/minute/1.73 m2 in the non-MNT group giving a significant mean difference of 12.5 and 12.8 when adjusted for linear regression (P , .001). Because the comparator group in this study was patients who had started dialysis in a 10-year period, we removed from the analyses patients who never started dialysis in a secondary analysis of the data. To adjust for the very sick, we also removed from the analysis those who started less than 1 year (,365 days) of diagnosis of CKD or date of MNT (Table 3, II. Patients at Stages 3, 4, 5 who started dialysis .365 days from the time of MNTor date of diagnosis). This adjustment reduced the sample size of patients who started dialysis to only 27 from the original 140 in the MNT group and 80 from the original 117 in the non-MNT group. At baseline, eGFR for all stages in patients who had started dialysis combined was 25.9 mL/minute/1.73 m2 for MNT versus 24.3 mL/minute/1.73 m2 for non-MNT group (2-sample t test P 5.5). From baseline to follow-up, the MNT group’s eGFR decreased less than the non-MNT group (decrease of 9.8 mL/minute/1.73 m2 for MNT group versus 12.3 mL/ minute/1.73 m2 for the non-MNT group). The difference of 2.5 was not significant (P 5 .37) but became significant when adjusted for stage (4.5, P 5 .04). Stratified by stage, the eGFR comparisons differed little for Stage 3 but for Stage 4, the MNT group decreased less than the nonMNT group (3.1 mL/minute/1.73 m2 vs. 10.1 mL/ minute/1.73 m2, P 5.05). Only 1 Stage 5 patient received MNT versus 14 in the non-MNT group, limiting the ability to make meaningful comparisons of Stage 5. A low eGFR was found for the high-risk group of patients that started dialysis within 1 year of diagnosis (Table 3, III. Patients at Stages 3, 4, 5 who started dialysis ,365 days from the time of MNTor date of diagnosis), but no significant differences was found between the groups. Time to the start of dialysis (Table 4, Fig. 1A) using survival analysis and Cox proportional hazards regression indicated that participants in the non-MNT group were 3.15
times more likely to initiate dialysis than the participants who received MNT. After adjusting for stage and baseline biomarkers, hazard ratio was 2.78 for the non-MNT group becoming more likely to initiate dialysis. Stratified by stage (Table 4, Fig. 1B and 1C), patients in the non-MNT group at Stage 3 and Stage 4 CKD had 3.47 and 3.45 more likelihood of starting dialysis, respectively. At Stage 5, the difference in probability of starting dialysis (Table 4, Fig. 1D) diminishes. The percent of biomarkers of CKD (calcium, phosphorus, iPTH, albumin, and BUN) within normal laboratory ranges between the MNTand non-MNT groups were higher in the MNT group at baseline and follow-up (Table 5). The albumin level was 4.6 times more likely to be within normal limits in the MNT group at follow-up compared with the non-MNT group, and when adjusted for stage, it was 4.24 times more likely as compared with the non-MNT group. At each individual stage (Stages 3, 4 and 5), albumin was more likely normal (5.37, 4.46, and 2.44, respectively) at follow-up in the MNT group. CKD-MBD biomarkers (calcium, phosphorus, iPTH) were more likely to be in normal range for the MNT group (3.51, 3.21, and 1.2, respectively) even when adjusted for stage (3.48, 3.07, and 1.22, respectively). The iPTH levels in the normal range were small for both groups (MNT vs. non-MNT), and no statistical significance was found between the groups. When biomarkers of CKD-MBD were stratified by stage (CKD Stages 3, 4, and 5), the levels were most likely to be in the normal ranges for Stages 3 and 4 for the MNT group, but that advantage diminishes for those seen in Stage 5. BUN levels within the normal range were very small for both the groups (MNT and nonMNT), and no statistical significance was observed. Stage-specific analysis was not possible for BUN levels because of the small number of patients in the normal range. Comparing only the patients who had started dialysis greater than 1 year from time of MNTor date of CKD diagnosis, the differences between the groups (MNT vs. nonMNT) were not significant (Table 6). However, the MNT groups had a higher percentage in the normal range for calcium, phosphorus, and albumin at follow-up. The non-MNT group had slightly higher percentages in the normal range for BUN and PTH, but the P value of 1 indicates that the observed differences could easily be
4
DE WAAL ET AL
Table 2. Participant Demographics and Baseline Characteristics by Quarter of Diagnosis Quarter of Entry into Analysis† Characteristic Group MNT Non-MNT Age (y) Mean MNT, mean (SD) Non-MNT, mean (SD) Gender Female MNT Non-MNT Male MNT Non-MNT Race Caucasian MNT Non-MNT Other§ MNT Non-MNT Stage CKD 2 MNT Non-MNT 3 MNT Non-MNT 4 MNT Non-MNT 5 MNT Non-MNT Diabetes Yes MNT Non-MNT No MNT Non-MNT
Total N
All
April 2003 to September 2005
October 2005 to March 2008
April 2008 to September 2010
October 2010 to April 2013
34 (23%) 5 (4%)
64 (44%) 56 (47%)
33 (22%) 45 (38%)
16 (11%) 12 (10%)
75 75 (13)
71 72 (12)
70 69 (11)
68 70 (10)
78 (10)
70 (15)
71 (11)
66 (19)
.39
17 (14%) 15 (21%) 2 (4%) 22 (16%) 19 (26%) 3 (5%)
59 (47%) 35 (48%) 24 (46%) 61 (44%) 29 (39%) 32 (48%)
34 (27%) 12 (16%) 22 (42%) 44 (31%) 21 (28%) 23 (35%)
15 (12%) 11 (15%) 4 (8%) 13 (9%) 5 (7%) 8 (12%)
.14
37 (15%) 32 (22%) 5 (5%) 2 (17%) 2 (50%) 0 (0%)
118 (47%) 64 (45%) 54 (49%) 2 (17%) 0 (0%) 2 (25%)
72 (28%) 31 (22%) 41 (37%) 6 (50%) 2 (50%) 4 (50%)
26 (10%) 16 (11%) 10 (9%) 2 (17%) 0 (0%) 2 (25%)
P Value*
147 118 265 147
71 72 (12)
118
70 (14)
125 73 52 140 74 66 253 143 110 12 4 8
.24
50% 44% 50% 56%
97% 93% 3% 7%
P Value‡ ,.001
.001jj 6 5 1 66 41 25 143 85 58 50 16 34 123 78 64 142 69 54
3% 1% 28% 21% 58% 49% 11% 29% .9 47% 46% 53% 54%
1 (17%) 1 (20%) 0 (0%) 10 (15%) 8 (20%) 2 (8%) 22 (15%) 21 (25%) 1 (2%) 6 (12%) 4 (25%) 2 (6%)
3 (50%) 2 (40%) 1 (100%) 36 (55%) 20 (49%) 16 (64%) 66 (46%) 36 (42%) 30 (52%) 15 (30%) 6 (38%) 9 (26%)
1 (17%) 1 (20%) 0 (0%) 14 (21%) 7 (17%) 7 (28%) 47 (33%) 22 (26%) 25 (43%) 16 (32%) 3 (19%) 13 (38%)
1 (17%) 1 (20%) 0 (0%) 6 (9%) 6 (15%) 0 (0%) 8 (6%) 6 (7%) 2 (3%) 13 (26%) 3 (19%) 10 (29%)
27 (22%) 10 (13%) 2 (3%) 12 (8%) 24 (35%) 3 (6%)
51 (41%) 34 (44%) 35 (55%) 69 (49%) 30 (43%) 21 (39%)
37 (30%) 21 (27%) 20 (31%) 41 (29%) 12 (17%) 25 (46%)
8 (7%) 13 (17%) 7 (11%) 20 (14%) 3 (4%) 5 (9%)
CKD, chronic kidney disease; MNT, medical nutrition therapy; SD, standard deviation. *Fisher’s exact test for categorical variables, 2-sample t test for continuous variables. Statistical significance P , .05. †Because care patterns likely changed over time, the number of patients in each group that came into the study were divided into the first, second, third, or fourth 2.5 years of the 10-year follow-up period. ‡P value comparing MNT versus Non-MNT across quarters (analysis of variance for age; Fisher’s exact test for categorical variables). §Other race: (N 5 MNT, non-MNT respectively): African American (2, 2), multiracial (2, 1), Hispanic (0, 1), Asian (0, 1), native American (0, 1), Hawaiian (0, 1), declined (0, 1). jj Global comparison of all 4 stages.
due to chance. Stratifying by stage, the sample sizes are small and lacked power to make any meaningful statistical assessment. For Stage 3, the differences were not significant, but the MNT group has much higher percentages in the normal range for calcium and albumin. On the other hand, MNT group had a lower percentage in the normal
range for phosphorus although not significant. In Stage 4, no differences were significant, but the MNT group had higher percentages in the normal range for all. Because only 1 Stage 5 patient started dialysis more than 1 year from time of MNT, comparisons were not performed for Stage 5 CKD.
5
MNT SLOWS PROGRESSION OF CKD, IMPROVES BIOMARKERS Table 3. Changes in eGFR Between Baseline and Follow-up Mean eGFR* Group
MNT
Non-MNT
Unadjusted Mean Difference
Adjusted† P Value
I. All patients Sample size 140 117 Baseline 28.0 21.7 6.3 ,.001 Follow-up 27.7 11.8 15.8 ,.001 Follow-up—baseline 20.3 29.9 9.6 ,.001 Patients in Stages 3 and 4 only Sample size 121 83 Baseline 28.5 26 2.5 .1 Follow-up 27.6 12.6 15 ,.001 Follow-up—baseline 21.0 213.5 12.5 ,.001 II. Patients at Stages 3, 4, 5 who started dialysis . 365 days from time of MNT or date of diagnosis All stages Sample size 27 80 Baseline 25.9 24.3 1.7 .50 Follow-up 16.1 11.9 4.2 .15 Follow-up—baseline 29.8 212.3 2.5 .37 Stage 3 at time of MNT or diagnosis Sample size 9 23 Baseline 41.6 37.4 4.2 .16 Follow-up 18.3 13.3 5.0 .43 Follow-up—baseline 223.2 224.0 0.8 .85 Stage 4 at time of MNT or diagnosis Sample size 17 43 Baseline 18.6 21.8 -3.2 ,.001 Follow-up 15.5 11.7 3.8 .26 Follow-up—baseline 23.1 210.1 7.0 .05 III. Patients at Stages 3, 4, 5 who started dialysis , 365 days from time of MNT or date of diagnosis All stages Sample size 14 36 Baseline 18.2 14.9 3.3 .23 Follow-up 13.4 11.3 2.1 .27 Follow-up—baseline 24.9 23.6 21.2 .69
Mean Difference
P Value
11.4
,.001
12.8
,.001
4.5
.04
21.7
.45
eGFR, estimated glomerular filtration rate; MNT, medical nutrition therapy. Follow-up is defined as date of dialysis initiation or the date of most recent laboratories if dialysis was not initiated. For P value, Fisher’s exact test for categorical variables, 2-sample t test for continuous variables. Statistical significance P , .05. Patients missing eGFR data in the MNT group were excluded in the analysis. *eGFR (mL/min/1.73 m2) was calculated as mL/min/1.73 m2 5 175 3 (serum creatinine) 2 1.154 3 (age) 2 0.203 3 (0.742 if female). †Adjusted for stage via linear regression.
Discussion The findings of this study suggest that the participants who received MNT from an RD with renal experience7,12,13 had better outcomes than participants who did not receive MNT. Participants who received MNT at Stages 3 and 4 CKD had better outcomes (slower progression to ESRD15 and better laboratory values) than the participants who received MNT at Stage 5 of CKD. Often, patients referred for MNT at Stage 5 are experiencing uremic symptoms, and MNT is used to minimize the symptoms. Nutritional status deteriorates as CKD progresses but is this due to uremia or poor dietary habits?16 Symptoms of uremia are usually the reasons patients start dialysis. High BUN levels are an indication of poor kidney health, so normal BUN levels are not expected at dialysis start and may be an indication that these patients were not eating well. Similar to this study, a previous study found
that participants with CKD who received MNTwere more likely to have normal albumin levels6 after the initiation of dialysis. Malnutrition and low serum albumins have a prognostic significance for outcomes in CKD and ESRD17,18 including hospitalization and mortality rates. To correct for the bias of the non-MNT group being patients who had started dialysis, we compared laboratory results of patients who started dialysis removing from the analysis patients who started dialysis less than 1 year from date of diagnosis or MNT as they may have been sicker and dialysis was more imminent. It is important to note that the MNT group was relatively small resulting in reduced statistical power. The fact that this group was so small may be because MNT slowed down the progression, and dialysis had not started as compared with a similar group that did not have MNT. Alternatively, a low eGFR was found to be the identifiable trait for those who started
6
DE WAAL ET AL
Table 4. Prediction of Starting Dialysis using Cox Proportional Hazards Regression Group
HR
95% Confidence P Interval Value
I. All patients A. Unadjusted Non-MNT versus MNT 3.58 2.54-5.05 ,.001 B. Adjusted for stage* Non-MNT versus MNT 3.15 2.20-4.49 ,.001 C. Adjusted for stage and baseline values of GFR, BUN, calcium, phosphorus, PTH, albumin* Non-MNT versus MNT 2.78 1.68-4.60 ,.001 II. Stratified by stage at time of MNT or diagnosis Stage 3 Non-MNT versus MNT 3.47 1.61-7.44 .001 Stage 4 Non-MNT versus MNT 3.45 2.19-5.44 ,.001 Stage 5 Non-MNT versus MNT 1.36 0.62-2.97 .44 BUN, blood urea nitrogen; GFR, glomerular filtration rate; HR, hazard ratio; MNT, medical nutrition therapy; PTH, parathyroid hormone. For P value, Fisher’s exact test for categorical variables, 2-sample t test for continuous variables for group differences. Statistical significance P , .05. *Adjusted via Cox proportional hazards regression.
dialysis within 1 year of diagnosis confirming a late referral to nephrology clinic.19 CKD-MBD progresses as CKD progresses. Reference ranges for iPTH are adjusted based on stage of CKD,3-5 so, a normal iPTH is not expected. Calcium and phosphorus levels are more indicative of better outcomes,20 and at follow-up, the MNT group had significantly better levels, but at dialysis start, it lacked the ability to derive significance due to small numbers. This study is consistent with results found in other studies demonstrating effectiveness of MNT for patients with CKD.7 This study had several limitations. First, all the participants in the non-MNT group were known to have started dialysis. We corrected for this bias by analyzing only patients who started dialysis but lacked power due to the small sample size. The demographics indicated a lack of diversity in the study population, reflective of the population base in this rural state.21 Because care patterns may have changed over time, we divided each group that came into the study by 2.5-year quarters of entry and found a statistical significance between the groups. The significance was more
Figure 1. Progression to dialysis (in days) of patients who received medical nutrition therapy (MNT) compared with those who did not. (A) At all stages of CKD at baseline. (B) Stratified by Stage 3 at baseline. (C) Stratified by Stage 4 at baseline. (D) Stratified by Stage 5. Time 0 is baseline, which is date of CKD diagnosis by a nephrologist and documented in the electronic medical records. Survival probability (Kaplan-Meier survival estimates) means the proportion of patients not on dialysis. (1 indicates that dialysis has not started, 0 indicates that dialysis had started). 1 Denotes participants who did not start dialysis during the duration of the study and were censored at the date of the most recent laboratory data.
Table 5. Percent of Study Group Within Normal Laboratory Ranges at Baseline and Follow-up
Laboratory Measure %Normal calcium GFR Stage 3 GFR Stage 4 GFR Stage 5 %Normal phosphorus GFR Stage 3 GFR Stage 4 GFR Stage 5 % Normal iPTH GFR Stage 3 GFR Stage 4 GFR Stage 5 % Normal Albumin GFR Stage 3 GFR Stage 4 GFR Stage 5 % Normal BUN
Follow-Up
OR (95% CI)
MNT
Non-MNT
P Value
MNT
Non-MNT
P Value
Unadjusted for Stage
Adjusted for Stage*
85.2 (n 5 142)
70.9 (n 5 117)
.006
75.69 (n 5 144)
47.01 (n 5 117)
,.0001
3.51 (2.07-5.94)
89.4 (n 5 133)
70.9 (n 5 117)
.0002
70.59 (n 5 144)
42.74 (n 5 117)
,.0001
3.21 (1.91-5.4)
26 (n 5 73)
20.8 (n 5 101)
.47
15.48 (n 5 84)
13.16 (n 5 114)
.68
1.2 (0.54-2.7)
85.7 (n 5 140)
69.9 (n 5 113)
.003
75.71 (n 5 84)
40.52 (n 5 116)
,.0001
4.6 (2.68-7.81)
5.5 (n 5 147)
2.54 (n 5 118)
.36
6.8 (n 5 147)
4.27 (n 5 117)
.43
1.64 (0.54-4.92)
3.48 (2.02-5.99) 8.9 (2.62-30.33) 3 (1.53-6.2) 1.9 (0.53-6.71) 3.07 (1.8-5.23) 6.36 (1.96-20.7) 3.46 (1.71-7.02) 1 (0.31-3.57) 1.22 (0.53-2.8) 1.34 (0.26-7.23) 1.21 (0.42-3.5) 1 (0.09-10.41) 4.24 (2.55-7.67) 5.37 (1.76-16.37) 4.46 (2.16-9.22) 2.44 (0.67-8.9) 0.92 (0.24-3.61)
BUN, blood urea nitrogen; CI, confidence interval; GFR, glomerular filtration rate; iPTH, intact parathyroid hormone; MNT, medical nutrition therapy; OR, odds ratio. For P value, Fisher’s exact test for categorical variables, 2-sample t test for continuous variables. Statistical significance P , .05. OR is the likelihood of biomarker to be within normal ranges. Follow-up is defined as date of dialysis initiation or the date of most recent laboratories if dialysis not initiated. For comparison purposes, normal laboratory values were considered between 8.5 and 10.1 mg/dL for calcium, 2.5 and 5.2 mg/dL for phosphorous, 3.4 and 5 g/dL for albumin, 7 and 18 mg/dL for BUN, and 14 and 72 pg/mL for PTH. *Adjusted for stage via logistic regression and stratification.
MNT SLOWS PROGRESSION OF CKD, IMPROVES BIOMARKERS
Baseline
7
8
DE WAAL ET AL
Table 6. Comparisons of Patients in Normal Range at Dialysis Start # (%) In Normal Range Group
MNT
All stages Sample size 29 BUN 1 (3%) Calcium 19 (68%) Phosphorous 13 (45%) PTH 2 (11%) Albumin 15 (58%) Stage 3 at time of MNT or diagnosis Sample size 9 BUN 0 (0%) Calcium 7 (78%) Phosphorous 3 (33%) PTH 0 (0%) Albumin 5 (71%) Stage 4 at time of MNT or diagnosis Sample size 19 BUN 1 (5%) Calcium 12 (67%) Phosphorous 9 (47%) PTH 2 (14%) Albumin 10 (56%)
Non-MNT
P Value
OR (95% CI)
80 3 (4%) 40 (50%) 32 (40%) 11 (14%) 32 (41%)
1.00 .13 .67 1.00 .17
0.92 (0.09-9.18) 2.11 (0.85-5.22) 1.22 (0.52-2.87) 0.71 (0.14-3.49) 2.00 (0.82-4.92)
23 0 (0%) 11 (48%) 11 (48%) 4 (18%) 8 (36%)
— .23 .69 1.00 .19
— 3.82 (0.65-22.45) 0.55 (0.11-2.73) — 4.38 (0.68-27.98)
43 2 (5%) 22 (51%) 14 (33%) 5 (12%) 21 (49%)
1.00 .40 .39 1.00 .78
1.14 (0.10-13.38) 1.91 (0.61-6.02) 1.86 (0.62-5.62) 1.23 (0.21-7.20) 1.31 (0.43-4.00)
BUN, blood urea nitrogen; CI, confidence interval; MNT, medical nutrition therapy; OR, odds ratio; PTH, parathyroid hormone. Analysis was restricted to patients who started dialysis greater than 1 year (.365 days) from date of MNT or date of diagnosis in Stages 3, 4, 5. For P value, Fisher’s exact test for categorical variables, 2-sample t test for continuous variables. Statistical significance P , .05. For comparison purposes, normal laboratory values were considered between 8.5 and 10.1 mg/dL for calcium, 2.5 and 5.2 mg/dL for phosphorous, 3.4 and 5 g/dL for albumin, 7 and 18 mg/dL for BUN, and 14 and 72 pg/mL for PTH.
related to sample size as there are no reasons to think that outcomes would be worse for patients seen in 2008 to 2010 than for patients seen in 2003 to 2005. The low sample size for the non-MNT group in the first 2.5-year quarter also may be due to this group had missing laboratory data and hence excluded from study. The electronic medical record began at our hospital in 2009, so, some of the earlier transcriptions and data were not available. Access to MNT for CKD in a rural state is difficult because of travel distance to the nephrology clinic. Number of MNT sessions with the RD was not differentiated, but each patient had at least 1 MNT session, which generally was between 60 and 90 minutes in length. Selection bias in this study may have little to do with gender, age, race, eGFR, diabetes, but rather things such as distance from center, support systems, which are surrogates for how well patients take care of themselves, follow medical advice, take all their medications, and frequency of follow-up visits. It could be argued that patients who do not seek MNT may also be nonadherent to medical therapies and other beneficial lifestyle factors, and the MNT group had more motivated patients who took the effort to seek MNT. This study did not assess whether the slope of the survival curve to dialysis changed after receiving MNT or control for other surrogate markers such as frequency of follow-up visits, medication compliance, distance from center, and other support systems.
Comorbidities and different CKD disease etiologies can impact disease progression rates. In a retrospective analysis, it was difficult to ascertain whether comorbidities were present at baseline or developed over time. Aging increases the incidence of comorbidities, so, adjustment for age partially corrects for the influence of comorbidity.22 In this study, the presence of diabetes but no other comorbidities was used in the analysis. Because diabetes is the major cause for CKD,23 the effect of other causes of CKD would be small. Controlling for stage of CKD in the analysis using adjustment and stratification should have mitigated differences in disease severity to some extent. There is a financial impact of reducing incidence of new ESRD. Medicare costs for CKD patients in 2012 was a mean of $20,162 per person year8,9 versus a mean of $70,000 per patient year for ESRD10 (dialysis and transplant) in 2009. In a simulation model for the future, burden of CKD in the United States has estimated continued significant growth in incidence of CKD based on current treatment patterns and risk factors.11 The Centers for Disease Control and Prevention document called Healthy People 202023 has identified CKD and ESRD as a significant public health problem and includes an objective to reduce the number of new cases of ESRD. The findings of this study suggest that MNT by an RD with renal expertise can slow down the progression to ESRD; hence, assisting in this objective. This study showed a delay in the initiation of dialysis but did not look at survival
MNT SLOWS PROGRESSION OF CKD, IMPROVES BIOMARKERS
once dialysis was initiated. Future research should not only assess MNT as one of the cost effective methods to slow the progression of CKD but also demonstrate an improvement on survival once dialysis was initiated. Currently, a new prospective nephrology clinic–based study by Arbor Research called the Chronic Kidney Disease Outcomes and Practice Patterns Study will be assessing best practice and outcomes with people with CKD.24 This study will be looking for evidence on treatment options, which improve QoL, delay dialysis and are cost effective. The database for this study started in 2014 in the United States. Accessing this database would be beneficial for comparisons of outcomes of patients who have had MNT and determine the characteristics (such as demographics, comorbidities, anthropometric measures, household income, QoL measures, insurance coverage, number of medications, nutrition-related biomarkers) and progression of CKD and compare with those who do not seek or have access to MNT.
Conclusions The evidence of this study suggests that MNT by an RD with renal expertise has a role in slowing the progression of CKD. MNT was associated with and may contribute to improved nutritional biomarkers. This has implications in decreasing CKD complications and reducing nutritional complications potentially improving QoL for people with CKD.
Practical Application MNT with an RD with renal experience should be encouraged when a patient is diagnosed with kidney disease. Improved nutrition awareness can contribute to a slowing in the progression of kidney disease and improved nutritional biomarkers in patients with CKD.
Acknowledgments The authors would like to thank the University of Vermont (UVM) Department of Nutrition and Food Science Department, Masters of Science in Dietetics (MSD) Program for this to be an Evidence-Based Research Project for an MSD student under the guidance of Amy Nickerson, MS, RD (Director of MSD Program at UVM). The authors would also like to acknowledge Cydney Wolf, MS, RD, CD (Clinical Dietitian Supervisor, Renal) and Virginia Hood, MBBS, MPH (Nephrologist), who reviewed and edited the article.
References 1. Centers for Medicare & Medicaid Services. National Coverage Determination (NCD) for Medical Nutrition Therapy (180.1). http://www.cms.gov/ Regulations-and-Guidance/Guidance/Transmittals/downloads/R154CIM. pdf. Accessed April 8, 2015. 2. National Kidney Foundation. K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: Evaluation, Classification and Stratification. Am J Kidney Dis. 2002;39(Suppl 1):S1-S266. 3. National Kidney Foundation. K/DOQI Clinical Practice Guidelines for Bone Metabolism and Disease in Chronic Kidney Disease. Am J Kidney Dis. 2003;42(Suppl 3):S1-S202. 4. Uhlig K, Berns JS, Kesterbaum B, et al. KDOQI US Commentary on the 2009 KDIGO Clinical Practice Guideline for the Diagnosis, Evaluation,
9
and Treatment of CKD-Mineral and Bone Disorder (CKD-MBD). Am J Kidney Dis. 2010;55:773-799. 5. Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2012 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease. Kidney Int. 2013;3:1-150. 6. Slinin Y, Guo H, Gilbertson D, Ensrud K, Collins A, Ishani A. Predialysis dietitian care & survival during the first year on dialysis. Am J Kidney Dis. 2011;58:583-590. 7. Academy of Nutrition and Dietetics Evidence Analysis Library. ‘‘ Is MNT provided by a Registered Dietitian for chronic kidney disease (CKD) in adults effective?’’ http://www.andeal.org/Accessed February 7, 2015. 8. Centers for Medicare & Medicaid Services, http://www.cms.gov/ Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/CMSStatistics-Reference-Booklet/index.html. Accessed February 2, 2015. 9. United States Renal Data System. 2014 Annual Data Report: Epidemiology of Kidney Disease in the United States. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases. http://www.usrds.org/. Accessed March 7, 2015. 10. National Kidney and Urologic Diseases Information Clearinghouse (NKUDIC), Kidney Disease Statistics for the United States. http://kidney. niddk.nih.gov/kudiseases/pubs/kustats/#17 Accessed February 2, 2015. 11. Hoerger TJ, Simpson SA, Yarnoff BO, et al. The Future Burden of CKD in the United States: A simulation Model for the CDC CKD Initiative. Am J Kidney Dis. 2015;65:403-411. 12. Kent PS, McCarthy MP, Burrowes JD, et al. Academy of Nutrition and Dietetics and National Kidney Foundation: Revised 2014 Standards of Practice and Standards of Professional Performance for Registered Dietitian Nutritionists (Competent, Proficient, and Expert) in Nephrology Nutrition. J Ren Nutr. 2014;24:275-285.e45. 13. Kent PS, McCarthy MP, Burrowes JD, et al. Academy of Nutrition and Dietetics and National Kidney Foundation: Revised 2014 Standards of Practice and Standards of Professional Performance for Registered Dietitian Nutritionists (Competent, Proficient, and Expert) in Nephrology Nutrition. J Acad Nutr Diet. 2014;114:1448-1457.e45. 14. Rothman KJ. No adjustments are needed for multiple comparisons. Epidemiology. 1990;1:43-46. 15. Levey AS, Inker LA, Matsushita K, et al. GFR Decline as an End Point for Clinical Trials in CKD: A Scientific Workshop Sponsored by the National Kidney Foundation and the US Food and Drug Administration. Am J Kidney Dis. 2014;64:821-835. 16. Chan M, Kelly J, Batterham M, Tapsell L. A High Prevalence of Abnormal Nutrition Parameter Found in Predialysis End-Stage Kidney Disease: Is It a Result of Uremia or Poor Eating Habits? J Ren Nutr. 2014;24:292-302. 17. Lawson JA, Lazarus R, Kelly JJ. Prevalence and Prognostic Significance of Malnutrition in Chronic Renal Insufficiency. J Ren Nutr. 2001;11:16-22. 18. Chan M, Kelly J, Batterham M, Tapsell L. Malnutrition (Subjective Global Assessment) Scores and Serum Albumin Levels, but not Body Mass Index Values, at Initiation of Dialysis are independent Predictors of Mortality: Q 10-Year Clinical Cohort Study. J Ren Nutr. 2012;22:547-557. 19. Wavamunno MD, Harris DCH. The need for early nephrology referral. Kidney Int. 2005;67(Suppl 94):S128-S132. 20. Floege J, Kim J, Ireland E, et al. Serum iPTH, calcium and phosphate, and the risk of mortality in a European Haemodialysis population. Nephrol Dial Transplant. 2011;26:1948-1955. 21. Vermont Quick Facts from the Census Bureau. http://quickfacts. census.gov/qfd/states/50000.html. Accessed June 13, 2015. 22. Van Manen JG, van Dijk PCW, Stel VS, et al. Confounding effect of comorbidity in survival studies in patients on renal replacement therapy. Nephrol Dial Transplant. 2007;22:187-195. 23. US Department of Health and Human Services. HealthyPeople.gov. Office of Disease Prevention and Health Promotion, Healthy People 2020.http://www.healthypeople.gov/ Accessed March 7, 2015. 24. Dialysis Outcomes and Practice Patterns Study Program. http://www. dopps.org/CKDOPPS/AbouttheCKDopps.aspx. Accessed March 7, 2015
