- Email: [email protected]
Geriatric Nutritional Risk Index as a Screening Tool for Malnutrition in Patients on Chronic Peritoneal Dialysis
Geriatric Nutritional Risk Index as a Screening Tool for Malnutrition in Patients on Chronic Peritoneal Dialysis Cheuk-Chun Szeto, MD, FRCP (Edin), Bonnie Ching-Ha Kwan, MBBS, MRCP (UK), Kai-Ming Chow, MBChB, MRCP (UK), Man-Ching Law, BN, RN, and Philip Kam-Tao Li, MD, FRCP Background: Malnutrition is common among peritoneal dialysis (PD) patients. Recently, the Geriatric Nutrition Risk Index (GNRI) was found to be a reliable tool for screening malnutrition in hemodialysis patients. However, the GNRI has not been validated in PD patients. Methods: We studied 314 unselected, adult PD patients from a single dialysis unit. We compared their GNRI scores with their comprehensive Malnutrition-Inflammation Scores (MIS) and 7-point Subjective Global Assessment (SGA) scores. We randomly selected 106 patients for a repeated assessment, and the changes in their three indices were compared. Results: Baseline GNRI was significantly correlated with MIS (r 5 20.487, P , .0001) and SGA (r 5 0.234, P , .0001). When MIS $6 was defined as malnutrition, the sensitivity and specificity of GNRI #93 in predicting malnutrition were 68.0% and 67.7%, respectively. When SGA #5 was used to define malnutrition, the sensitivity and specificity were 54.5% and 71.1%, respectively. The change in GNRI was correlated with the change in MIS (r 5 20.244, P 5 .012) and overall SGA score (r 5 0.266, P 5 .006), respectively. When an increase in MIS was defined as a worsening of nutrition, the sensitivity and specificity of GNRI were 45.7% and 81.7%, respectively. When a decrease in SGA was used to define a worsening of nutrition, the sensitivity and specificity were 42.3% and 87.0%, respectively. Conclusions: Although GNRI is significantly correlated with other nutritional indices, it is not sensitive for screening malnutrition in PD patients. Serial measurements of GNRI are also not sensitive in detecting a change in nutritional status. Further study is needed to identify a simple and reliable tool for the assessment and monitoring of nutritional status in PD patients. Ó 2010 by the National Kidney Foundation, Inc. All rights reserved.
M
ALNUTRITION IS A common and serious problem in patients with endstage renal disease treated with peritoneal dialysis (PD).1–3 Malnutrition in PD patients may be caused by a variety of factors, including anorexia, predialysis dietary restriction, altered protein metabolism, dialysis-related nutrient losses, acidosis, and inflammation.4,5 More importantly, malnourished patients have significantly increased mortality and morbidity, compared with their well-nourished Department of Medicine and Therapeutics, Prince of Wales Hospital, Chinese University of Hong Kong, Shatin, Hong Kong, China. Address reprint requests to Cheuk-Chun Szeto, MD, FRCP (Edin), Department of Medicine and Therapeutics, Prince of Wales Hospital, Chinese University of Hong Kong, Shatin, Hong Kong, China. E-mail: [email protected] Ó 2010 by the National Kidney Foundation, Inc. All rights reserved. 1051-2276/10/2001-0005$36.00/0 doi:10.1053/j.jrn.2009.04.004
Journal of Renal Nutrition, Vol 20, No 1 (January), 2010: pp 29–37
counterparts.6 Because the development of malnutrition is a gradual process, a simple and reliable method of assessment that could be conducted frequently is required for monitoring PD patients. However, an optimal technique for the nutritional assessment of PD patient remains to be established. The Malnutrition Inflammation Score (MIS) was recently developed.7 The MIS system incorporates the seven components of the Subjective Global Assessment (SGA), together with body mass index, serum albumin level, and total iron-binding capacity (TIBC), in an attempt to achieve a comprehensive assessment of nutritional status. The MIS was shown to have significant associations with prospective hospitalization and mortality in patients on maintenance hemodialysis, and it was superior to its individual components as well as conventional SGA for predicting mortality.7 Our previous study also showed that MIS has a reasonable correlation with conventional SGA scores in PD patients.8 29
30
SZETO ET AL
On the other hand, some simpler nutritional screening tools can be scored according to welldefined rules or cutoffs. These tools involve common measures, and can be applied rapidly in a large population. Notably, the Geriatric Nutritional Risk Index (GNRI) was designed and validated by Bouillanne et al.9 for screening malnutrition in elderly medical patients. Yamada et al.10 found that the GNRI was the simplest and most accurate index for identifying hemodialysis patients at nutritional risk according to the MIS. The characteristics of the SGA, MIS, and GNRI as screening tool of malnutrition are summarized in Table 1. However, the GNRI has not been validated in PD patients. Here, we report on a cohort study examining the applicability of the GNRI to PD patients. In addition, we examined the role of the GNRI, compared with the MIS, in the serial monitoring of nutritional status in PD patients.
PD to undergo a repeated assessment. In each case, consent was obtained. This study was approved by the Clinical Research Ethics Committee of the Chinese University of Hong Kong. The overall design of the study is summarized in Figure 1.
Patients and Methods
Data Collection Baseline clinical data, including age, sex, duration of dialysis, PD regimen, body weight and height, underlying renal disease, and presence of comorbid conditions, were obtained by chart review. The modified Charlson’s Comorbidity Index, as validated in PD patients,11 was used to calculate a comorbidity score. Data on dialysis adequacy indices, including total Kt/V, normalized protein nitrogen appearance (NPNA), fat-free edema-free body mass (FEBM), and residual glomerular filtration rate, as determined by previously described methods,12,13 were retrieved from clinical records. All patients were followed until July 1, 2008, and their survival status was recorded.
Patient Selection From May 2006 to May 2007, we studied 314 unselected, adult (age over 18 years) PD patients from the Renal Unit of the Prince of Wales Hospital, Shatin, Hong Kong. Nine patients received automated PD, whereas all others received continuous ambulatory peritoneal dialysis. After 1 year of initial assessment, 34 patients died, 7 were converted to chronic hemodialysis, 6 underwent a kidney transplant, and 5 were transferred to other centers. We randomly selected 106 patients who remained on
Assessment of Nutritional Status Nutritional status was assessed by SGA and MIS. For the SGA, the four-item, 7-point system was used.13,14 The four items for assessment included changes in body weight, degree of anorexia, amount of subcutaneous tissue, and muscle mass. The four individual-item scores were combined to generate a global score, which also took into account the clinical judgment of observers and thus did not represent a simple arithmetic aggregate of
Table 1. Characteristics of Subjective Global Assessment, Malnutrition Inflammation Score, and Geriatric Nutritional Risk Index
Components Weight loss/BMI Dietary intake Gastrointestinal symptoms Comorbidity Physical function Signs of muscle/fat loss Serum albumin Total iron-binding capacity Complexity Sensitivity to detect change Correlation with clinical outcome Validated in PD patients
Subjective Global Assessment
Subjective Global Assessment
Geriatric Nutritional Risk Index
Yes Yes No No No Yes No No Simple Poor Good Yes
Yes Yes Yes Yes Yes Yes Yes Yes Complex Good Good Yes
Yes No No No No No Yes No Simple Unknown Unknown No
BMI, body mass index; PD, peritoneal dialysis.
GERIATRIC NUTRITIONAL RISK INDEX FOR PD Figure 1. Overall design of study. GNRI, Geriatric Nutritional Risk Index; MIS, Malnutrition-Inflammation Score; SGA, Subjective Global Assessment; PD, peritoneal dialysis.
31
314 unselected adult PD patients nutritional assessment: SGA, MIS, GNRI
one year later
106 patients (randomly selected)
the four individual-item scores. All SGA items were rated subjectively on a scale from 1 to 7, where 1 or 2 represents severe malnutrition, 3 to 5 represent moderate to mild malnutrition, and 6 or 7 represents mild malnutrition to normal nutritional status.14 Our previous study showed that the Cohen’s k concordant coefficient for agreement between observers was 0.84,13 i.e., an excellent level of agreement. Calculation of the MIS was described previously.7,8 Briefly, the MIS is a score that incorporates the SGA and three other components. The MIS consists of four main parts: a patient’s related medical history, physical examination, body mass index, and laboratory parameters. A patient’s medical history includes weight changes, dietary intake, gastrointestinal symptoms, functional capacity, and comorbidity, including number of years on dialysis. The physical examination detects any loss of subcutaneous fat and signs of muscle wasting. Laboratory parameters include serum albumin and serum TIBC levels. Serum albumin was measured by the bromcresol purple method. The 10 components were scored from 0 (normal) to 3 (very severe), and thus a total score ranges from 0 to 30.
change in nutritional status
2nd nutritional assessment: SGA, MIS, GNRI
Body-weight-to-ideal-body-weight ratio was set to 1 when a patient’s body weight exceeded the ideal body weight. The ideal body weight in the present study was calculated by the formula of Lorentz,9,15 as in the original GNRI equation.
Statistical Analysis Statistical analysis was performed using SPSS for Windows, version 11.5 (SPSS, Inc., Chicago, IL). All data are expressed as mean 6 SD, unless otherwise specified. Data were compared according to the Mann-Whitney U-test or Spearman’s rankcorrelation coefficient, as appropriate. P , 0.05 was taken as statistically significant. All probabilities were two-tailed. In addition, with the use of the MIS or SGA as the reference standard, receiver operating characteristic (ROC) curves were generated for the GNRI. The area under the ROC curve (AUC) indicated the probability of discriminating a nutritional risk. The cutoff risk point of nutrition for each reference standard was then defined from the highest sensitivity (1 – specificity) value in the ROC curve.
Results Geriatric Nutritional Risk Index The GNRI was developed by modifying the nutritional risk index for elderly patients. This index is calculated based on serum albumin and body weight, using the following equation:9,10 GNRI5½1:4893albuminðg=dLÞ1½41:7 3ðbody weight=ideal body weightÞ
We studied 314 PD patients. The baseline demographic and clinical data are summarized in Table 2. Of these patients, 250 (79.6%) had 3 exchanges per day, whereas the others had 4 exchanges per day. Thirteen patients (4.1%) were underweight by World Health Organization standards, with a body mass index below 18.5 kg/m2. When MIS was considered as the reference standard, 47 patients (15.0%) had moderate
32
SZETO ET AL
Table 2. Baseline Demographic and Clinical Data Number of patients
314
Sex (male:female) 149:165 Age (years) 60.0 6 11.9 Duration of dialysis (months) 39.0 6 38.6 Body height (cm) 158.9 6 8.1 Body weight (kg) 61.7 6 12.3 24.4 6 4.1 Body mass index (kg/m2) Blood pressure (mm Hg) Systolic 146 6 21 Diastolic 77 6 13 Renal diagnosis, number of cases (%) Glomerulonephritis 96 (30.6%) Diabetic nephropathy 101 (32.2%) Polycystic kidney 7 (2.2%) Hypertensive nephrosclerosis 28 (8.9%) Obstructive uropathy 16 (5.1%) Others/unknown 66 (21.0%) Major comorbidity, number of cases (%) Diabetes mellitus 130 (41.1%) Ischemic heart disease 46 (14.6%) Cerebrovascular disease 42 (13.4%) Peripheral vascular disease 15 (4.8%) Charlson’s Comorbidity Index 5.42 6 2.27 Dialysis adequacy Total Kt/V 2.00 6 0.47 1.98 6 2.24 Residual GFR (mL/min/1.73 m2) NPNA (g/kg/day) 1.09 6 0.24 FEBM (%) 52.4 6 14.1 GFR, glomerular filtration rate; NPNA, normalized protein nitrogen appearance; FEBM, fat-free edema-free body mass. All values 6 SD.
to severe malnutrition (MIS $11), and 142 (45.2%) had mild malnutrition (MIS 6 to 10). When overall SGA score was considered as the reference standard, 3 patients (1.0%) had moderate to severe malnutrition (SGA #3), and 87 (27.7%) had mild malnutrition (SGA 4 to 5).
Relationship With Other Nutritional Indices The correlation between the GNRI and baseline biochemical parameters, dialysis adequacy, and nutritional indices are summarized in Table 3. In brief, the GNRI had a significant correlation with MIS, overall SGA score, NPNA, and Charlson’s Comorbidity Index, but only a marginal correlation with FEBM, and no correlation with total Kt/V. When MIS $6 was defined as malnutrition, we explored the accuracy of GNRI for a diagnosis of malnutrition. The ROC curve, as shown in Figure 2, has an AUC of 0.737, indicating that GNRI is moderately useful for the diagnosis of
Table 3. Correlation Between Geriatric Nutritional Risk Index and Baseline Biochemical Parameters, Dialysis Adequacy, and Nutritional Indices Parameter Age Duration of dialysis Body weight Body mass index Mean blood pressure Serum albumin Hemoglobin Lipid profile Total cholesterol Triglyceride LDL cholesterol HDL cholesterol MIS Weight change Dietary intake Gastrointestinal symptoms Functional capacity Comorbid conditions Subcutaneous fat Muscle wasting Body mass index score Albumin score TIBC score Total score Charlson’s Comorbidity Index SGA score Body weight Anorexia Subcutaneous fat Muscle mass Overall score Kt/V Residual GFR NPNA FEBM
Correlation Coefficient*
P Value
r 5 20.117 r 5 20.012 r 5 0.256 r 5 0.307 r 5 0.052 r 5 0.941 r 5 0.246
.038 .8 ,.0001 ,.0001 .4 ,.0001 ,.0001
r 5 0.053 r 5 0.114 r 5 0.038 r 5 20.147
.4 .051 .5 .009
r 5 20.117 r 5 20.066 r 5 20.107
.038 .24 .058
r 5 20.097 r 5 20.130 r 5 20.227 r 5 20.262 r 5 20.313
.09 .021 ,.0001 ,.0001 ,.0001
r 5 20.651 r 5 20.337 r 5 20.487 r 5 20.210
,.0001 ,.0001 ,.0001 .0002
r 5 0.160 r 5 0.087 r 5 0.223 r 5 0.235 r 5 0.234 r 5 0.090 r 5 0.196 r 5 0.257 r 5 20.111
.005 .13 .0001 ,.0001 ,.0001 .12 .002 .0001 .08
LDL, low-density lipoprotein; HDL, high-density lipoprotein; MIS, Malnutrition-Inflammation Score; TIBC, total iron-binding capacity; SGA, Subjective Global Assessment; GFR, glomerular filtration rate; NPNA, normalized protein nitrogen appearance; FEBM, fat-free edema-free body mass. *Spearman’s rank correlation coefficient.
malnutrition. The GNRI’s sensitivity and specificity of #93 in predicting malnutrition according to the MIS were 68.0% and 67.7%, respectively. When an overall SGA score #5 was defined as malnutrition, the AUC of the ROC curve of the GNRI for a diagnosis of malnutrition was 0.643 (Fig. 2). The GNRI’s sensitivity and specificity
33
GERIATRIC NUTRITIONAL RISK INDEX FOR PD
A
B
1.0
0.8
sensitivity
sensitivity
0.8
1.0
0.6
0.4
0.2
0.6
0.4
0.2
AUC = 0.737
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1 - specificity
AUC = 0.643
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1 - specificity
Figure 2. Receiver operating characteristic curve of Geriatric Nutritional Risk Index for a diagnosis of malnutrition, with (A) Malnutrition-Inflammation Score $6 and (B) overall Subjective Global Assessment score $5 as reference standards of malnutrition. AUC, area under the curve.
of #93 in predicting malnutrition according to overall SGA score were 54.5% and 71.1%, respectively.
Changes in Nutrition Status In 106 patients who were followed for 1 year, the MIS increased (indicating a deterioration in nutritional status) in 60 (56.6%), it remained static in 19 (17.9%), and it decreased in 27 (25.5%). During this period, the overall SGA score decreased in 54 (50.9%), remained static in 37 (34.9%), and increased in 15 (14.2%). The change in GNRI had a modest but statistically significant correlation with the change in MIS (r 5 20.244, P 5 .012) and overall SGA score (r 5 0.266, P 5.006). The reliability of change in GNRI for a diagnosis of change in nutritional status was explored by ROC curves, as shown in Figure 3. The sensitivity and specificity of GNRI for a diagnosis of change in nutritional status (using change in MIS or overall SGA score as reference standard) are summarized in Table 4. In general, a change in GNRI has reasonable specificity but limited sensitivity in detecting a change in nutritional status. Relationship With Survival As of July 1, 2008, 49 (15.6%) of the original 314 patients had died. Compared with the survivors, patients who died during the observation period had a lower overall SGA score (5.27 6 0.91 vs. 5.94 6 0.82, P ,.0001), higher MIS (9.53 6 3.79 vs. 6.25 6 3.79, P ,.0001), and lower GNRI
(87.1 6 7.4 vs. 92.9 6 6.7, P ,.0001), all indicating worse nutritional status.
Discussion In the present study, we found that in Chinese PD patients, the GNRI had a statistically significant correlation with other indices of nutritional status, such as MIS and SGA. However, the correlation was modest. The GNRI is not sufficiently sensitive to serve as a screening tool for malnutrition in Chinese PD patients. The ideal marker of nutritional status in PD patients remains controversial.16,17 Serum chemistry, body mass, muscle mass, dietary intake, energy expenditure, body-composition assessment, and various nutritional scoring systems have all been advocated.16 Among these, serum albumin is often used because its effect on patient outcomes is widely accepted.12,14 The National Kidney Foundation of the United States recommended the use of SGA, NPNA, and FEBM as nutritional indices for PD patients.18 In theory, anthropometric measurements and assessments of body composition (e.g., by bioelectric impedance) are attractive approaches to nutritional assessment. However, these methods could prove difficult for PD patients because of their alterations in water distribution and body composition.19,20 Our previous study showed that lean body mass, as determined by anthropometric methods, correlates poorly with other markers of nutritional status.21
34
SZETO ET AL
B
MIS for worsening of nutrition
MIS for improvement of nutrition
1.0
1.0
0.8
0.8
sensitivity
sensitivity
A
0.6
0.4
0.6
0.4
0.2
0.2 AUC = 0.616
0.0 0.0
0.2
0.4
0.6
0.8
AUC = 0.567
0.0 0.0
1.0
0.2
1 - specificity
D
SGA for worsening of nutrition
0.6
0.8
1.0
SGA for improvement of nutrition
1.0
1.0
0.8
0.8
sensitivity
sensitivity
C
0.4
1 - specificity
0.6
0.4
0.6
0.4
0.2
0.2 AUC = 0.660
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1 - specificity
AUC = 0.682
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1 - specificity
Figure 3. Receiver operating characteristic curve of change in Geriatric Nutritional Risk Index for a diagnosis of change in nutritional status. (A) Worsening of nutrition, as defined by an increase in Malnutrition-Inflammation Score (MIS). (B) Improvement in nutrition, as defined by a decrease in MIS. (C) Worsening of nutrition, as defined by a decrease in overall Subjective Global Assessment (SGA) score. (D) Improvement in nutrition, as defined by an increase in overall SGA score. AUC, area under the curve.
To the best of our knowledge, this is the first study to test the use of GNRI in the screening of malnutrition in PD patients. Our results are different from those of Yamada et al.,10 who reported that the GNRI is an accurate tool for the identification of malnourished hemodialysis patients. In that study, the sensitivity and specificity, and accuracy, of GNRI at a cutoff of 91.2, in predicting malnutrition according to the MIS, were 0.730 and 0.819, respectively.10 In addition to modality of dialysis, there are important differences in the patient population between the study by Yamada et al.10 and the present study. Notably, patients were dialyzed much longer in the former report
(mean duration, .10 years),10 whereas our patients were dialyzed for an average of around 3 years (Table 2). The patients of Yamada et al.10 had a much lower body mass index (averaging around 19 kg/m2), compared with an average of around 24 kg/m2 in our report. The difference in patients’ body mass indices between the study of Yamada et al.10 and our present study may explain the difference in the sensitivity and specificity of GNRI for the predication of malnutrition between the two studies. It is noteworthy that a measured body weight (or its difference from ideal body weight) plays an important part in the calculation of GNRI, and
35
GERIATRIC NUTRITIONAL RISK INDEX FOR PD Table 4. Change in Geriatric Nutritional Risk Index for a Diagnosis of Change in Nutritional Status Worsening of Nutrition
Improvement in Nutrition
Reference Standard
MIS Increase
SGA Decrease
MIS Decrease
SGA Increase
Geriatric Nutritional Risk Index Cutoff* Sensitivity Specificity
0 45.7% 81.7%
1.38 42.3% 87.0%
0.63 44.4% 74.7%
1.58 53.3% 81.3%
MIS, Malnutrition-Inflammation Score; SGA, overall Subjective Global Assessment score. *Cutoff values represent change in Geriatric Nutritional Risk Index identified by receiver operating characteristic curve (see Fig. 3), and have optimal sensitivity and specificity.
in patients with a high body mass index, GNRI becomes dependent only on serum albumin level, making it an insensitive marker of malnutrition. We are not sure, however, whether the higher body mass indices in our patient population, compared with the chronic hemodialysis patients of Yamada et al.,10 were the result of a higher prevalence of obesity, (subclinical) fluid overload, or both. Overweight status and chronic volume overload are both recognized as common problems in chronic PD patients.22–24 It could be argued that the GNRI was originally designed for the identification of malnutrition in a geriatric population,9 and that extrapolations to young dialysis patients may not be appropriate. However, our results remained similar when only patients aged over 65 years were analyzed (details not shown). Irrespective of the underlying reason, our study highlights the uniqueness of PD patients, because GNRI has been previously validated in hemodialysis.10 One strength of our present study is its inclusion of a cohort of over 100 patients with a repeated assessment of nutritional status 1 year later, so that we could examine the accuracy of GNRI as a tool for the detection of changes in nutritional status. In essence, we found that a change in GNRI has a reasonable specificity but poor sensitivity (below 50%) in identifying a change in nutritional status, when a change in MIS or overall SGA score was used as the reference standard (Table 4). Although the numeric results appear satisfactory, and GNRI is simple to calculate, we believe that serial measurements of GNRI would add little to the routine clinical care of PD patients. The GNRI is meant as a screening tool for the identification of malnutrition (or change in nutritional status); high sensitivity is the desirable characteristic, and there is no need for high specificity. In line with Buzby
et al.25 and Bouillanne et al.,9 we set the bodyweight ratio to 1 when the measured body weight exceeded the ideal one. This methodology, however, is controversial. A decrease in body-weight ratio could have been missed in overweight or overhydrated patients whose body ratio had been initially set at 1, and a change in GNRI may have been missed in the longitudinal study. The choice of cutoff value for the identification of malnutrition according to SGA and MIS is arbitrary. An SGA score of #5 was taken to indicate malnutrition according to current recommendations of the National Kidney Foundation.26 An MIS score of $6 was taken to indicate malnutrition, because a previous study showed that the risk of peritonitis rose sharply above this cutoff.27 Based on the above cutoff values for MIS, however, up to 60% of PD patients had some degree of malnutrition, whereas the prevalence was around 30% when based on SGA. Our recent studies suggest that MIS $8 (compared with 6) may have a better correlation with SGA in terms of identifying malnutrition.8,28 With this definition of MIS, 36% of our PD population had malnutrition, a figure similar to that obtained using the SGA. It is important to realize that no simple scoring system of nutritional status has been validated for longitudinal measurements. In the present study, we used MIS and overall SGA score as the reference standard for changes in nutritional status, because these two parameters are the best-reported in the literature. Although SGA has not been formally validated for longitudinal measurements, the Canada-USA Study showed that SGAwas related to the mortality and hospitalization of PD patients when the score was considered as a time-dependent variable.14 Our previous study also suggested that the SGA might be used for monitoring responses to changes in therapeutic trials.13 Similarly, although
36
SZETO ET AL
the MIS has only been validated as a descriptive and predictive variable,7,8 its individual components are widely used in the serial monitoring of dialysis patients. Because of its detailed scoring system, the MIS should, at least theoretically, be a more sensitive tool compared with the SGA in detecting a change in nutritional status using serial measurements. Further research in this area should focus on the validation of a simple screening tool for the presence of malnutrition, as well as the identification of a reliable instrument for the serial measurement and detection of changes in the nutritional status of PD patients. Because of budgetary constraints, we randomly selected 106 patients from our original cohort for repeated assessment after 1 year, and so our study may not have adequate statistical power to detect small but clinically meaningful changes in nutritional status.
Conclusion Although the GNRI is significantly correlated with other nutritional indices, it is not sensitive for screening malnutrition in PD patients. Serial measurements of the GNRI were also not sensitive in detecting changes of nutritional status in PD patients. Further studies are needed to identify a simple, reliable tool for the assessment and monitoring of nutritional status in PD patients.
Acknowledgments This study was supported in part by Chinese University of Hong Kong research account 6901031 and the Richard Yu Chinese University of Hong Kong Peritoneal Dialysis Research Fund. The authors declare no conflicts of interest.
References 1. Fenton SS, Johnston N, Delmore T, et al: Nutritional assessment of continuous ambulatory peritoneal dialysis patients. ASAIO Trans 33:650-653, 1987 2. Kaminski MV Jr, Lowrie EG, Rosenblatt SG, Haase T: Malnutrition is lethal, diagnosable and treatable in ESRD patients. Transplant Proc 23:1810-1815, 1991 3. Szeto CC, Chow KM: Metabolic acidosis and malnutrition in dialysis patients. Semin Dial 17:371-375, 2004 4. Bossola M, Muscaritoli M, Tazza L, et al: Malnutrition in hemodialysis patients: what therapy? Am J Kidney Dis 46: 371-386, 2005 5. Stenvinkel P, Heimburger O, Lindholm B, Kaysen GA, Bergstrom J: Are there two types of malnutrition in chronic renal failure? Evidence for relationships between malnutrition, inflammation and atherosclerosis (MIA syndrome). Nephrol Dial Transplant 15:953-960, 2000 6. Davies SJ, Phillips L, Naish PF, Russell GI: Quantifying comorbidity in peritoneal dialysis patients and its relationship to
other predictors of survival. Nephrol Dial Transplant 17: 1085-1092, 2002 7. Kalantar-Zadeh K, Kopple JD, Block G, Humphreys MH: A malnutrition-inflammation score is correlated with morbidity and mortality in maintenance hemodialysis patients. Am J Kidney Dis 38:1251-1263, 2001 8. Chan JY, Che KI, Lam KM, et al: Comprehensive Malnutrition Inflammation Score as a marker of nutritional status in Chinese peritoneal dialysis patients. Nephrology 12:130-134, 2007 9. Bouillanne O, Morineau G, Dupont C, et al: Geriatric Nutritional Risk Index: a new index for evaluating at-risk elderly medical patients. Am J Clin Nutr 82:777-783, 2005 10. Yamada K, Furuya R, Takita T, et al: Simplified nutritional screening tools for patients on maintenance hemodialysis. Am J Clin Nutr 87:106-113, 2008 11. Beddhu S, Zeidel ML, Saul M, et al: The effects of comorbid conditions on the outcomes of patients undergoing peritoneal dialysis. Am J Med 112:696-701, 2002 12. Szeto CC, Wong TY, Leung CB, et al: Importance of dialysis adequacy in mortality and morbidity of Chinese CAPD patients. Kidney Int 58:400-407, 2000 13. Szeto CC, Wong TY, Chow KM, Leung CB, Li PK: Oral sodium bicarbonate for the treatment of metabolic acidosis in peritoneal dialysis patients—a randomized placebo-control trial. J Am Soc Nephrol 14:2119-2126, 2003 14. Canada-USA (CANUSA) Peritoneal Dialysis Study Group: Adequacy of dialysis and nutrition in continuous peritoneal dialysis: association with clinical outcomes. J Am Soc Nephrol 7:198-207, 1996 15. Lorentz FH: Der Konstitutionsindex der Frau. Klin Wochenshr 16:734-736, 1929 16. Fouque D, Kalantar-Zadeh K, Kopple J, et al: A proposed nomenclature and diagnostic criteria for protein-energy wasting in acute and chronic kidney disease. Kidney Int 73: 391-398, 2008 17. Chung SH, Stenvinkel P, Lindholm B, Avesani CM: Identifying and managing malnutrition stemming from different causes. Perit Dial Int 27(Suppl 2):S239-S244, 2007 18. Peritoneal Dialysis Adequacy Work Group: NKF-K/ DOQI clinical practice guidelines for peritoneal dialysis adequacy: update 2000. Am J Kidney Dis 37(Suppl 1):S65-S136, 2000 19. Szeto CC, Lai KN, Wong TY, Law MC, Li PK: Measuredto-predicted creatinine generation ratio increases with time and decline in residual renal function in continuous ambulatory peritoneal dialysis. Am J Kidney Dis 34:235-241, 1999 20. Woodrow G, Devine Y, Cullen M, Lindley E: Application of bioelectrical impedance to clinical assessment of body composition in peritoneal dialysis. Perit Dial Int 27:496-502, 2007 21. Szeto CC, Kong J, Wu AK, Wong TY, Wang AY, Li PK: The role of lean body mass as nutritional index in Chinese peritoneal dialysis patients—comparison of creatinine kinetic and anthropometric methods. Perit Dial Int 20:708-714, 2000 22. Li PK, Kwan BC, Szeto CC, Ko GT: Metabolic syndrome in peritoneal dialysis patients. Nephrol Dial Transplant Plus 4: 206-214, 2008 23. Lameire N, Van Biesen W: Importance of blood pressure and volume control in peritoneal dialysis patients. Perit Dial Int 21:206-211, 2001 24. Ates K, Nergizoglu G, Keven K, et al: Effect of fluid and sodium removal on mortality in peritoneal dialysis patients. Kidney Int 60:767-776, 2001
GERIATRIC NUTRITIONAL RISK INDEX FOR PD 25. Buzby GP, Knox LS, Crosby LO, et al: Study protocol: a randomized clinical trial of total parenteral nutrition in malnourished surgical patients. Am J Clin Nutr 47(Suppl 2): 366-381, 1988 26. Peritoneal Dialysis Adequacy Work Group: Clinical practice guidelines for peritoneal dialysis adequacy. Am J Kidney Dis 48(Suppl 1):S98-S129, 2006
37
27. Afsar B, Sezer S, Ozdemir FN, Celik H, Elsurer R, Haberal M: Malnutrition-Inflammation Score is a useful tool in peritoneal dialysis patients. Perit Dial Int 26:705-711, 2006 28. Cheng TM, Lam DH, Ting SK, et al: Serial monitoring of nutritional status in Chinese peritoneal dialysis patients by Subjective Global Assessment and comprehensive Malnutrition Inflammation Score. Nephrology 14:143-174, 2009
M
ALNUTRITION IS A common and serious problem in patients with endstage renal disease treated with peritoneal dialysis (PD).1–3 Malnutrition in PD patients may be caused by a variety of factors, including anorexia, predialysis dietary restriction, altered protein metabolism, dialysis-related nutrient losses, acidosis, and inflammation.4,5 More importantly, malnourished patients have significantly increased mortality and morbidity, compared with their well-nourished Department of Medicine and Therapeutics, Prince of Wales Hospital, Chinese University of Hong Kong, Shatin, Hong Kong, China. Address reprint requests to Cheuk-Chun Szeto, MD, FRCP (Edin), Department of Medicine and Therapeutics, Prince of Wales Hospital, Chinese University of Hong Kong, Shatin, Hong Kong, China. E-mail: [email protected] Ó 2010 by the National Kidney Foundation, Inc. All rights reserved. 1051-2276/10/2001-0005$36.00/0 doi:10.1053/j.jrn.2009.04.004
Journal of Renal Nutrition, Vol 20, No 1 (January), 2010: pp 29–37
counterparts.6 Because the development of malnutrition is a gradual process, a simple and reliable method of assessment that could be conducted frequently is required for monitoring PD patients. However, an optimal technique for the nutritional assessment of PD patient remains to be established. The Malnutrition Inflammation Score (MIS) was recently developed.7 The MIS system incorporates the seven components of the Subjective Global Assessment (SGA), together with body mass index, serum albumin level, and total iron-binding capacity (TIBC), in an attempt to achieve a comprehensive assessment of nutritional status. The MIS was shown to have significant associations with prospective hospitalization and mortality in patients on maintenance hemodialysis, and it was superior to its individual components as well as conventional SGA for predicting mortality.7 Our previous study also showed that MIS has a reasonable correlation with conventional SGA scores in PD patients.8 29
30
SZETO ET AL
On the other hand, some simpler nutritional screening tools can be scored according to welldefined rules or cutoffs. These tools involve common measures, and can be applied rapidly in a large population. Notably, the Geriatric Nutritional Risk Index (GNRI) was designed and validated by Bouillanne et al.9 for screening malnutrition in elderly medical patients. Yamada et al.10 found that the GNRI was the simplest and most accurate index for identifying hemodialysis patients at nutritional risk according to the MIS. The characteristics of the SGA, MIS, and GNRI as screening tool of malnutrition are summarized in Table 1. However, the GNRI has not been validated in PD patients. Here, we report on a cohort study examining the applicability of the GNRI to PD patients. In addition, we examined the role of the GNRI, compared with the MIS, in the serial monitoring of nutritional status in PD patients.
PD to undergo a repeated assessment. In each case, consent was obtained. This study was approved by the Clinical Research Ethics Committee of the Chinese University of Hong Kong. The overall design of the study is summarized in Figure 1.
Patients and Methods
Data Collection Baseline clinical data, including age, sex, duration of dialysis, PD regimen, body weight and height, underlying renal disease, and presence of comorbid conditions, were obtained by chart review. The modified Charlson’s Comorbidity Index, as validated in PD patients,11 was used to calculate a comorbidity score. Data on dialysis adequacy indices, including total Kt/V, normalized protein nitrogen appearance (NPNA), fat-free edema-free body mass (FEBM), and residual glomerular filtration rate, as determined by previously described methods,12,13 were retrieved from clinical records. All patients were followed until July 1, 2008, and their survival status was recorded.
Patient Selection From May 2006 to May 2007, we studied 314 unselected, adult (age over 18 years) PD patients from the Renal Unit of the Prince of Wales Hospital, Shatin, Hong Kong. Nine patients received automated PD, whereas all others received continuous ambulatory peritoneal dialysis. After 1 year of initial assessment, 34 patients died, 7 were converted to chronic hemodialysis, 6 underwent a kidney transplant, and 5 were transferred to other centers. We randomly selected 106 patients who remained on
Assessment of Nutritional Status Nutritional status was assessed by SGA and MIS. For the SGA, the four-item, 7-point system was used.13,14 The four items for assessment included changes in body weight, degree of anorexia, amount of subcutaneous tissue, and muscle mass. The four individual-item scores were combined to generate a global score, which also took into account the clinical judgment of observers and thus did not represent a simple arithmetic aggregate of
Table 1. Characteristics of Subjective Global Assessment, Malnutrition Inflammation Score, and Geriatric Nutritional Risk Index
Components Weight loss/BMI Dietary intake Gastrointestinal symptoms Comorbidity Physical function Signs of muscle/fat loss Serum albumin Total iron-binding capacity Complexity Sensitivity to detect change Correlation with clinical outcome Validated in PD patients
Subjective Global Assessment
Subjective Global Assessment
Geriatric Nutritional Risk Index
Yes Yes No No No Yes No No Simple Poor Good Yes
Yes Yes Yes Yes Yes Yes Yes Yes Complex Good Good Yes
Yes No No No No No Yes No Simple Unknown Unknown No
BMI, body mass index; PD, peritoneal dialysis.
GERIATRIC NUTRITIONAL RISK INDEX FOR PD Figure 1. Overall design of study. GNRI, Geriatric Nutritional Risk Index; MIS, Malnutrition-Inflammation Score; SGA, Subjective Global Assessment; PD, peritoneal dialysis.
31
314 unselected adult PD patients nutritional assessment: SGA, MIS, GNRI
one year later
106 patients (randomly selected)
the four individual-item scores. All SGA items were rated subjectively on a scale from 1 to 7, where 1 or 2 represents severe malnutrition, 3 to 5 represent moderate to mild malnutrition, and 6 or 7 represents mild malnutrition to normal nutritional status.14 Our previous study showed that the Cohen’s k concordant coefficient for agreement between observers was 0.84,13 i.e., an excellent level of agreement. Calculation of the MIS was described previously.7,8 Briefly, the MIS is a score that incorporates the SGA and three other components. The MIS consists of four main parts: a patient’s related medical history, physical examination, body mass index, and laboratory parameters. A patient’s medical history includes weight changes, dietary intake, gastrointestinal symptoms, functional capacity, and comorbidity, including number of years on dialysis. The physical examination detects any loss of subcutaneous fat and signs of muscle wasting. Laboratory parameters include serum albumin and serum TIBC levels. Serum albumin was measured by the bromcresol purple method. The 10 components were scored from 0 (normal) to 3 (very severe), and thus a total score ranges from 0 to 30.
change in nutritional status
2nd nutritional assessment: SGA, MIS, GNRI
Body-weight-to-ideal-body-weight ratio was set to 1 when a patient’s body weight exceeded the ideal body weight. The ideal body weight in the present study was calculated by the formula of Lorentz,9,15 as in the original GNRI equation.
Statistical Analysis Statistical analysis was performed using SPSS for Windows, version 11.5 (SPSS, Inc., Chicago, IL). All data are expressed as mean 6 SD, unless otherwise specified. Data were compared according to the Mann-Whitney U-test or Spearman’s rankcorrelation coefficient, as appropriate. P , 0.05 was taken as statistically significant. All probabilities were two-tailed. In addition, with the use of the MIS or SGA as the reference standard, receiver operating characteristic (ROC) curves were generated for the GNRI. The area under the ROC curve (AUC) indicated the probability of discriminating a nutritional risk. The cutoff risk point of nutrition for each reference standard was then defined from the highest sensitivity (1 – specificity) value in the ROC curve.
Results Geriatric Nutritional Risk Index The GNRI was developed by modifying the nutritional risk index for elderly patients. This index is calculated based on serum albumin and body weight, using the following equation:9,10 GNRI5½1:4893albuminðg=dLÞ1½41:7 3ðbody weight=ideal body weightÞ
We studied 314 PD patients. The baseline demographic and clinical data are summarized in Table 2. Of these patients, 250 (79.6%) had 3 exchanges per day, whereas the others had 4 exchanges per day. Thirteen patients (4.1%) were underweight by World Health Organization standards, with a body mass index below 18.5 kg/m2. When MIS was considered as the reference standard, 47 patients (15.0%) had moderate
32
SZETO ET AL
Table 2. Baseline Demographic and Clinical Data Number of patients
314
Sex (male:female) 149:165 Age (years) 60.0 6 11.9 Duration of dialysis (months) 39.0 6 38.6 Body height (cm) 158.9 6 8.1 Body weight (kg) 61.7 6 12.3 24.4 6 4.1 Body mass index (kg/m2) Blood pressure (mm Hg) Systolic 146 6 21 Diastolic 77 6 13 Renal diagnosis, number of cases (%) Glomerulonephritis 96 (30.6%) Diabetic nephropathy 101 (32.2%) Polycystic kidney 7 (2.2%) Hypertensive nephrosclerosis 28 (8.9%) Obstructive uropathy 16 (5.1%) Others/unknown 66 (21.0%) Major comorbidity, number of cases (%) Diabetes mellitus 130 (41.1%) Ischemic heart disease 46 (14.6%) Cerebrovascular disease 42 (13.4%) Peripheral vascular disease 15 (4.8%) Charlson’s Comorbidity Index 5.42 6 2.27 Dialysis adequacy Total Kt/V 2.00 6 0.47 1.98 6 2.24 Residual GFR (mL/min/1.73 m2) NPNA (g/kg/day) 1.09 6 0.24 FEBM (%) 52.4 6 14.1 GFR, glomerular filtration rate; NPNA, normalized protein nitrogen appearance; FEBM, fat-free edema-free body mass. All values 6 SD.
to severe malnutrition (MIS $11), and 142 (45.2%) had mild malnutrition (MIS 6 to 10). When overall SGA score was considered as the reference standard, 3 patients (1.0%) had moderate to severe malnutrition (SGA #3), and 87 (27.7%) had mild malnutrition (SGA 4 to 5).
Relationship With Other Nutritional Indices The correlation between the GNRI and baseline biochemical parameters, dialysis adequacy, and nutritional indices are summarized in Table 3. In brief, the GNRI had a significant correlation with MIS, overall SGA score, NPNA, and Charlson’s Comorbidity Index, but only a marginal correlation with FEBM, and no correlation with total Kt/V. When MIS $6 was defined as malnutrition, we explored the accuracy of GNRI for a diagnosis of malnutrition. The ROC curve, as shown in Figure 2, has an AUC of 0.737, indicating that GNRI is moderately useful for the diagnosis of
Table 3. Correlation Between Geriatric Nutritional Risk Index and Baseline Biochemical Parameters, Dialysis Adequacy, and Nutritional Indices Parameter Age Duration of dialysis Body weight Body mass index Mean blood pressure Serum albumin Hemoglobin Lipid profile Total cholesterol Triglyceride LDL cholesterol HDL cholesterol MIS Weight change Dietary intake Gastrointestinal symptoms Functional capacity Comorbid conditions Subcutaneous fat Muscle wasting Body mass index score Albumin score TIBC score Total score Charlson’s Comorbidity Index SGA score Body weight Anorexia Subcutaneous fat Muscle mass Overall score Kt/V Residual GFR NPNA FEBM
Correlation Coefficient*
P Value
r 5 20.117 r 5 20.012 r 5 0.256 r 5 0.307 r 5 0.052 r 5 0.941 r 5 0.246
.038 .8 ,.0001 ,.0001 .4 ,.0001 ,.0001
r 5 0.053 r 5 0.114 r 5 0.038 r 5 20.147
.4 .051 .5 .009
r 5 20.117 r 5 20.066 r 5 20.107
.038 .24 .058
r 5 20.097 r 5 20.130 r 5 20.227 r 5 20.262 r 5 20.313
.09 .021 ,.0001 ,.0001 ,.0001
r 5 20.651 r 5 20.337 r 5 20.487 r 5 20.210
,.0001 ,.0001 ,.0001 .0002
r 5 0.160 r 5 0.087 r 5 0.223 r 5 0.235 r 5 0.234 r 5 0.090 r 5 0.196 r 5 0.257 r 5 20.111
.005 .13 .0001 ,.0001 ,.0001 .12 .002 .0001 .08
LDL, low-density lipoprotein; HDL, high-density lipoprotein; MIS, Malnutrition-Inflammation Score; TIBC, total iron-binding capacity; SGA, Subjective Global Assessment; GFR, glomerular filtration rate; NPNA, normalized protein nitrogen appearance; FEBM, fat-free edema-free body mass. *Spearman’s rank correlation coefficient.
malnutrition. The GNRI’s sensitivity and specificity of #93 in predicting malnutrition according to the MIS were 68.0% and 67.7%, respectively. When an overall SGA score #5 was defined as malnutrition, the AUC of the ROC curve of the GNRI for a diagnosis of malnutrition was 0.643 (Fig. 2). The GNRI’s sensitivity and specificity
33
GERIATRIC NUTRITIONAL RISK INDEX FOR PD
A
B
1.0
0.8
sensitivity
sensitivity
0.8
1.0
0.6
0.4
0.2
0.6
0.4
0.2
AUC = 0.737
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1 - specificity
AUC = 0.643
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1 - specificity
Figure 2. Receiver operating characteristic curve of Geriatric Nutritional Risk Index for a diagnosis of malnutrition, with (A) Malnutrition-Inflammation Score $6 and (B) overall Subjective Global Assessment score $5 as reference standards of malnutrition. AUC, area under the curve.
of #93 in predicting malnutrition according to overall SGA score were 54.5% and 71.1%, respectively.
Changes in Nutrition Status In 106 patients who were followed for 1 year, the MIS increased (indicating a deterioration in nutritional status) in 60 (56.6%), it remained static in 19 (17.9%), and it decreased in 27 (25.5%). During this period, the overall SGA score decreased in 54 (50.9%), remained static in 37 (34.9%), and increased in 15 (14.2%). The change in GNRI had a modest but statistically significant correlation with the change in MIS (r 5 20.244, P 5 .012) and overall SGA score (r 5 0.266, P 5.006). The reliability of change in GNRI for a diagnosis of change in nutritional status was explored by ROC curves, as shown in Figure 3. The sensitivity and specificity of GNRI for a diagnosis of change in nutritional status (using change in MIS or overall SGA score as reference standard) are summarized in Table 4. In general, a change in GNRI has reasonable specificity but limited sensitivity in detecting a change in nutritional status. Relationship With Survival As of July 1, 2008, 49 (15.6%) of the original 314 patients had died. Compared with the survivors, patients who died during the observation period had a lower overall SGA score (5.27 6 0.91 vs. 5.94 6 0.82, P ,.0001), higher MIS (9.53 6 3.79 vs. 6.25 6 3.79, P ,.0001), and lower GNRI
(87.1 6 7.4 vs. 92.9 6 6.7, P ,.0001), all indicating worse nutritional status.
Discussion In the present study, we found that in Chinese PD patients, the GNRI had a statistically significant correlation with other indices of nutritional status, such as MIS and SGA. However, the correlation was modest. The GNRI is not sufficiently sensitive to serve as a screening tool for malnutrition in Chinese PD patients. The ideal marker of nutritional status in PD patients remains controversial.16,17 Serum chemistry, body mass, muscle mass, dietary intake, energy expenditure, body-composition assessment, and various nutritional scoring systems have all been advocated.16 Among these, serum albumin is often used because its effect on patient outcomes is widely accepted.12,14 The National Kidney Foundation of the United States recommended the use of SGA, NPNA, and FEBM as nutritional indices for PD patients.18 In theory, anthropometric measurements and assessments of body composition (e.g., by bioelectric impedance) are attractive approaches to nutritional assessment. However, these methods could prove difficult for PD patients because of their alterations in water distribution and body composition.19,20 Our previous study showed that lean body mass, as determined by anthropometric methods, correlates poorly with other markers of nutritional status.21
34
SZETO ET AL
B
MIS for worsening of nutrition
MIS for improvement of nutrition
1.0
1.0
0.8
0.8
sensitivity
sensitivity
A
0.6
0.4
0.6
0.4
0.2
0.2 AUC = 0.616
0.0 0.0
0.2
0.4
0.6
0.8
AUC = 0.567
0.0 0.0
1.0
0.2
1 - specificity
D
SGA for worsening of nutrition
0.6
0.8
1.0
SGA for improvement of nutrition
1.0
1.0
0.8
0.8
sensitivity
sensitivity
C
0.4
1 - specificity
0.6
0.4
0.6
0.4
0.2
0.2 AUC = 0.660
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1 - specificity
AUC = 0.682
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1 - specificity
Figure 3. Receiver operating characteristic curve of change in Geriatric Nutritional Risk Index for a diagnosis of change in nutritional status. (A) Worsening of nutrition, as defined by an increase in Malnutrition-Inflammation Score (MIS). (B) Improvement in nutrition, as defined by a decrease in MIS. (C) Worsening of nutrition, as defined by a decrease in overall Subjective Global Assessment (SGA) score. (D) Improvement in nutrition, as defined by an increase in overall SGA score. AUC, area under the curve.
To the best of our knowledge, this is the first study to test the use of GNRI in the screening of malnutrition in PD patients. Our results are different from those of Yamada et al.,10 who reported that the GNRI is an accurate tool for the identification of malnourished hemodialysis patients. In that study, the sensitivity and specificity, and accuracy, of GNRI at a cutoff of 91.2, in predicting malnutrition according to the MIS, were 0.730 and 0.819, respectively.10 In addition to modality of dialysis, there are important differences in the patient population between the study by Yamada et al.10 and the present study. Notably, patients were dialyzed much longer in the former report
(mean duration, .10 years),10 whereas our patients were dialyzed for an average of around 3 years (Table 2). The patients of Yamada et al.10 had a much lower body mass index (averaging around 19 kg/m2), compared with an average of around 24 kg/m2 in our report. The difference in patients’ body mass indices between the study of Yamada et al.10 and our present study may explain the difference in the sensitivity and specificity of GNRI for the predication of malnutrition between the two studies. It is noteworthy that a measured body weight (or its difference from ideal body weight) plays an important part in the calculation of GNRI, and
35
GERIATRIC NUTRITIONAL RISK INDEX FOR PD Table 4. Change in Geriatric Nutritional Risk Index for a Diagnosis of Change in Nutritional Status Worsening of Nutrition
Improvement in Nutrition
Reference Standard
MIS Increase
SGA Decrease
MIS Decrease
SGA Increase
Geriatric Nutritional Risk Index Cutoff* Sensitivity Specificity
0 45.7% 81.7%
1.38 42.3% 87.0%
0.63 44.4% 74.7%
1.58 53.3% 81.3%
MIS, Malnutrition-Inflammation Score; SGA, overall Subjective Global Assessment score. *Cutoff values represent change in Geriatric Nutritional Risk Index identified by receiver operating characteristic curve (see Fig. 3), and have optimal sensitivity and specificity.
in patients with a high body mass index, GNRI becomes dependent only on serum albumin level, making it an insensitive marker of malnutrition. We are not sure, however, whether the higher body mass indices in our patient population, compared with the chronic hemodialysis patients of Yamada et al.,10 were the result of a higher prevalence of obesity, (subclinical) fluid overload, or both. Overweight status and chronic volume overload are both recognized as common problems in chronic PD patients.22–24 It could be argued that the GNRI was originally designed for the identification of malnutrition in a geriatric population,9 and that extrapolations to young dialysis patients may not be appropriate. However, our results remained similar when only patients aged over 65 years were analyzed (details not shown). Irrespective of the underlying reason, our study highlights the uniqueness of PD patients, because GNRI has been previously validated in hemodialysis.10 One strength of our present study is its inclusion of a cohort of over 100 patients with a repeated assessment of nutritional status 1 year later, so that we could examine the accuracy of GNRI as a tool for the detection of changes in nutritional status. In essence, we found that a change in GNRI has a reasonable specificity but poor sensitivity (below 50%) in identifying a change in nutritional status, when a change in MIS or overall SGA score was used as the reference standard (Table 4). Although the numeric results appear satisfactory, and GNRI is simple to calculate, we believe that serial measurements of GNRI would add little to the routine clinical care of PD patients. The GNRI is meant as a screening tool for the identification of malnutrition (or change in nutritional status); high sensitivity is the desirable characteristic, and there is no need for high specificity. In line with Buzby
et al.25 and Bouillanne et al.,9 we set the bodyweight ratio to 1 when the measured body weight exceeded the ideal one. This methodology, however, is controversial. A decrease in body-weight ratio could have been missed in overweight or overhydrated patients whose body ratio had been initially set at 1, and a change in GNRI may have been missed in the longitudinal study. The choice of cutoff value for the identification of malnutrition according to SGA and MIS is arbitrary. An SGA score of #5 was taken to indicate malnutrition according to current recommendations of the National Kidney Foundation.26 An MIS score of $6 was taken to indicate malnutrition, because a previous study showed that the risk of peritonitis rose sharply above this cutoff.27 Based on the above cutoff values for MIS, however, up to 60% of PD patients had some degree of malnutrition, whereas the prevalence was around 30% when based on SGA. Our recent studies suggest that MIS $8 (compared with 6) may have a better correlation with SGA in terms of identifying malnutrition.8,28 With this definition of MIS, 36% of our PD population had malnutrition, a figure similar to that obtained using the SGA. It is important to realize that no simple scoring system of nutritional status has been validated for longitudinal measurements. In the present study, we used MIS and overall SGA score as the reference standard for changes in nutritional status, because these two parameters are the best-reported in the literature. Although SGA has not been formally validated for longitudinal measurements, the Canada-USA Study showed that SGAwas related to the mortality and hospitalization of PD patients when the score was considered as a time-dependent variable.14 Our previous study also suggested that the SGA might be used for monitoring responses to changes in therapeutic trials.13 Similarly, although
36
SZETO ET AL
the MIS has only been validated as a descriptive and predictive variable,7,8 its individual components are widely used in the serial monitoring of dialysis patients. Because of its detailed scoring system, the MIS should, at least theoretically, be a more sensitive tool compared with the SGA in detecting a change in nutritional status using serial measurements. Further research in this area should focus on the validation of a simple screening tool for the presence of malnutrition, as well as the identification of a reliable instrument for the serial measurement and detection of changes in the nutritional status of PD patients. Because of budgetary constraints, we randomly selected 106 patients from our original cohort for repeated assessment after 1 year, and so our study may not have adequate statistical power to detect small but clinically meaningful changes in nutritional status.
Conclusion Although the GNRI is significantly correlated with other nutritional indices, it is not sensitive for screening malnutrition in PD patients. Serial measurements of the GNRI were also not sensitive in detecting changes of nutritional status in PD patients. Further studies are needed to identify a simple, reliable tool for the assessment and monitoring of nutritional status in PD patients.
Acknowledgments This study was supported in part by Chinese University of Hong Kong research account 6901031 and the Richard Yu Chinese University of Hong Kong Peritoneal Dialysis Research Fund. The authors declare no conflicts of interest.
References 1. Fenton SS, Johnston N, Delmore T, et al: Nutritional assessment of continuous ambulatory peritoneal dialysis patients. ASAIO Trans 33:650-653, 1987 2. Kaminski MV Jr, Lowrie EG, Rosenblatt SG, Haase T: Malnutrition is lethal, diagnosable and treatable in ESRD patients. Transplant Proc 23:1810-1815, 1991 3. Szeto CC, Chow KM: Metabolic acidosis and malnutrition in dialysis patients. Semin Dial 17:371-375, 2004 4. Bossola M, Muscaritoli M, Tazza L, et al: Malnutrition in hemodialysis patients: what therapy? Am J Kidney Dis 46: 371-386, 2005 5. Stenvinkel P, Heimburger O, Lindholm B, Kaysen GA, Bergstrom J: Are there two types of malnutrition in chronic renal failure? Evidence for relationships between malnutrition, inflammation and atherosclerosis (MIA syndrome). Nephrol Dial Transplant 15:953-960, 2000 6. Davies SJ, Phillips L, Naish PF, Russell GI: Quantifying comorbidity in peritoneal dialysis patients and its relationship to
other predictors of survival. Nephrol Dial Transplant 17: 1085-1092, 2002 7. Kalantar-Zadeh K, Kopple JD, Block G, Humphreys MH: A malnutrition-inflammation score is correlated with morbidity and mortality in maintenance hemodialysis patients. Am J Kidney Dis 38:1251-1263, 2001 8. Chan JY, Che KI, Lam KM, et al: Comprehensive Malnutrition Inflammation Score as a marker of nutritional status in Chinese peritoneal dialysis patients. Nephrology 12:130-134, 2007 9. Bouillanne O, Morineau G, Dupont C, et al: Geriatric Nutritional Risk Index: a new index for evaluating at-risk elderly medical patients. Am J Clin Nutr 82:777-783, 2005 10. Yamada K, Furuya R, Takita T, et al: Simplified nutritional screening tools for patients on maintenance hemodialysis. Am J Clin Nutr 87:106-113, 2008 11. Beddhu S, Zeidel ML, Saul M, et al: The effects of comorbid conditions on the outcomes of patients undergoing peritoneal dialysis. Am J Med 112:696-701, 2002 12. Szeto CC, Wong TY, Leung CB, et al: Importance of dialysis adequacy in mortality and morbidity of Chinese CAPD patients. Kidney Int 58:400-407, 2000 13. Szeto CC, Wong TY, Chow KM, Leung CB, Li PK: Oral sodium bicarbonate for the treatment of metabolic acidosis in peritoneal dialysis patients—a randomized placebo-control trial. J Am Soc Nephrol 14:2119-2126, 2003 14. Canada-USA (CANUSA) Peritoneal Dialysis Study Group: Adequacy of dialysis and nutrition in continuous peritoneal dialysis: association with clinical outcomes. J Am Soc Nephrol 7:198-207, 1996 15. Lorentz FH: Der Konstitutionsindex der Frau. Klin Wochenshr 16:734-736, 1929 16. Fouque D, Kalantar-Zadeh K, Kopple J, et al: A proposed nomenclature and diagnostic criteria for protein-energy wasting in acute and chronic kidney disease. Kidney Int 73: 391-398, 2008 17. Chung SH, Stenvinkel P, Lindholm B, Avesani CM: Identifying and managing malnutrition stemming from different causes. Perit Dial Int 27(Suppl 2):S239-S244, 2007 18. Peritoneal Dialysis Adequacy Work Group: NKF-K/ DOQI clinical practice guidelines for peritoneal dialysis adequacy: update 2000. Am J Kidney Dis 37(Suppl 1):S65-S136, 2000 19. Szeto CC, Lai KN, Wong TY, Law MC, Li PK: Measuredto-predicted creatinine generation ratio increases with time and decline in residual renal function in continuous ambulatory peritoneal dialysis. Am J Kidney Dis 34:235-241, 1999 20. Woodrow G, Devine Y, Cullen M, Lindley E: Application of bioelectrical impedance to clinical assessment of body composition in peritoneal dialysis. Perit Dial Int 27:496-502, 2007 21. Szeto CC, Kong J, Wu AK, Wong TY, Wang AY, Li PK: The role of lean body mass as nutritional index in Chinese peritoneal dialysis patients—comparison of creatinine kinetic and anthropometric methods. Perit Dial Int 20:708-714, 2000 22. Li PK, Kwan BC, Szeto CC, Ko GT: Metabolic syndrome in peritoneal dialysis patients. Nephrol Dial Transplant Plus 4: 206-214, 2008 23. Lameire N, Van Biesen W: Importance of blood pressure and volume control in peritoneal dialysis patients. Perit Dial Int 21:206-211, 2001 24. Ates K, Nergizoglu G, Keven K, et al: Effect of fluid and sodium removal on mortality in peritoneal dialysis patients. Kidney Int 60:767-776, 2001
GERIATRIC NUTRITIONAL RISK INDEX FOR PD 25. Buzby GP, Knox LS, Crosby LO, et al: Study protocol: a randomized clinical trial of total parenteral nutrition in malnourished surgical patients. Am J Clin Nutr 47(Suppl 2): 366-381, 1988 26. Peritoneal Dialysis Adequacy Work Group: Clinical practice guidelines for peritoneal dialysis adequacy. Am J Kidney Dis 48(Suppl 1):S98-S129, 2006
37
27. Afsar B, Sezer S, Ozdemir FN, Celik H, Elsurer R, Haberal M: Malnutrition-Inflammation Score is a useful tool in peritoneal dialysis patients. Perit Dial Int 26:705-711, 2006 28. Cheng TM, Lam DH, Ting SK, et al: Serial monitoring of nutritional status in Chinese peritoneal dialysis patients by Subjective Global Assessment and comprehensive Malnutrition Inflammation Score. Nephrology 14:143-174, 2009