- Email: [email protected]
A comparison of two methods of dietary assessment in peritoneal dialysis patients
A Comparison of Two Methods of Dietary Assessment in Peritoneal Dialysis Patients Anne Gr@ths, BSc, SRD, Lesley Russell, BSc, SRD, Mavion Breslin, BSc, SRD, Gavin Russell, MD, FRO, and Simon Davies, MD, FRCP
Objective: To conduct a comparison of two methods of dietary assessment in patients on peritoneal dialysis. Design: Comparative, cross-sectional study of two methods of dietary assessment (3-day diet diary and 24-hour recall). Data was collected simultaneously by a single experienced dietitian. Each assessment was coded and analyzed blind. Setting: Regional speciality peritoneal dialysis training unit. Patients: In this study, 30 peritoneal dialysis patients recruited prospectively and consecutively as they attended for out-patient assessment of dialysis adequacy. Age range was 22 to 77 years. Patients were excluded if unwell, younger than 18 years, or had peritonitis. Outcome measures: Total energy and protein intakes from both methods were compared. Protein intakes from both methods were compared with the protein catabolic rate generated from urea kinetics. Data obtained from both methods were compared using paired ttests, linear regression, and Bland and Altman techniques. Results: There were no significant differences in the mean daily protein (72.4 g ~76.6 g) and total energy (1757 kcals (7.35 MJ) v. 1897 kcals (7.94 MJ)) intakes determined by the two methods. Positive correlations were seen between the measurements for protein intake (r = .58, P = .0026) energy intake (r = .78, P < .OOOOl), with mean differences of .066 g/kg/d (SD .38) 2.04 kcal/kg/day (SD 6.67), respectively. For both methods there was a similar positive correlation between dietary protein intake and protein catabolic rate. A 24-hour recall was more likely to result in successful collection of data (29 of 30) than 3-day diet diaries (25 of 30) and was less time consuming. Conclusion: These two methods of determining dietary protein and energy intake do not differ significantly in the information they provide. The relative success in obtaining completed records of intake, the shorter time taken and the opportunity for patient education and assessment of other nutrition related factors has led to the adoption of the 24-hour recall method in our institution. o 1999 by the National Kidney Foundation, inc.
M
ALNUTRITION is a well documented and frequent occurrence in patients on peritoneal dialysis1 in whom it has been shown to be a predictor of poor outcome2 and to be associated with a higher incidence of mortality and morbidity.3 For these reasons it is important that there is regular dietary assessment in peritoneal dialysis patients, so that those at risk can be identified and early intervention initiated. There are many methods of dietary assessment: these include 7-day weighed intakes, photographic records, 3-day patient-kept diet diaries
0 1999 by the Natiowal Kidney Fouddioiz, 1051.2276/99/0901-0005$3.00/0
26
Inc.
using household measures, diet histories, and 24-hour dietary recall. 4,5 Each has advantages and disadvantages. In the clinical setting, the method of dietary assessment needs to be quick and accurate and yield a good response rate from the patients.6 In the absence of a gold standard method of dietary assessment, 3-day food diaries are commonly used to assess dietary intake.’ They involve a patient being given a diary, writing down a description of all food and drink consumed during a 3-day period using household measures (eg, 1 tablespoon or 1 cup, etc), instead of using weighing scales. This is less restrictive for the patient than weighing each item of food and hopefully will mean that there is greater compliance from the patient to complete the diary in detail. The potential advantages are two-fold; firstly it is thought that less time is needed by the dietitian to obtain the information and secondly that patients Jmrnal
ofRenal
Nutrition,
Vol9,
No 1 (January),
1999: pp 26-31
COMPARISON
OF TWO
METHODS
do not have to rely on their memories to complete it. However, it needs motivation and compliance from the patient, and often insufficient detail about food and drink is recorded. In addition and importantly, the patient needs to be literate. Some of the problems associated with 3-day diet diaries can be avoided by using 24-hour dietary recall. This involves a dietitian asking a patient to recall their food and drink intake within the past 24 hours. Numerous questions are asked to prompt the patient and aid memory. The dietitian can also clarify at the time of the interview the portion sizes eaten, and an accurate impression of all food and drink consumed can then be made. However, this method of dietary assessment does not take into account day-to-day variations of food and drink intake and patients with poor memories may not be able to recall their intake in enough detail to enable an accurate assessment. However, in reality, a peritoneal dialysis patient’s lifestyle is punctuated by their treatment regimen, which may, as a result, lead to less variation in day-to-day dietary intake (unpublished observations). If this were the case, then 24 hour recall may give a reasonable estimate of a dialysis patients intake. As an alternative to dietetic assessment, protein intake may be calculated using urea kinetics. Estimates of the protein catabolic rate (PCR) derived from the 24-hour urea nitrogen generation and protein losses. In carefully conducted nitrogen balance studies, this has been found to correlate with the dietary protein intake. However, this has only been shown to be the case in stable patients, where there is no urea generated from muscle breakdown.’ One of the disadvantages of this approach to assessing adequate dietary intake in peritoneal dialysis patients is its failure to assess total calorie intake. The aim of this study was to conduct a comparison of 24-hour dietary recall and 3-day diet diaries using household measures to establish if these methods are equivalent in their reliability as tools for dietary assessment in peritoneal dialysis patients. To validate these dietary assessments of protein intake both methods were also compared with PCR obtained from urea kinetics in patients whose body weight was stable as judged by a stable mid-arm circumference.9
OF DlETARY
27
ASSESSMENT
Patients and Methods Patient Population Thirty patients treated by peritoneal dialysis were recruited prospectively and consecutively as they attended for outpatient assessment of dialysis adequacy. Patients were excluded if they had an acute intercurrent illness, were younger than 18 years of age, or had peritonitis within the last month. The purpose of this selection was to obtain a stable dialysis population.
Methods of Dietary Assessment: Day Diet Diaries
Three
Before assessment, each patient completed a 3-day diet diary at home. This involved the patient filling in the diary in detail to include all food and drink consumed during the 3 days immediately preceding the assessment of dialysis adequacy. This may or may not have included a weekend day. The patients were asked to describe the food items and portion size in as much detail as possible using household measures. Food diaries also contained a short questionnaire to improve the detail of written diaries, which asked about food items that patients frequently forget to detail in food diaries hence prompting memory (eg, type of spread on bread, sugar in drinks, type and size of bread loaf). A brief written explanation on how to complete 3-day diet diaries was also included, and this was explained to them by the dietitian.
24-Hour
Dietary
Recall
A single experienced dietitian performed a 24-hour recall for the day before assessment of dialysis adequacy. This involves precise questioning of the subject of the previous days food and drink intake. The recall begins from the first food or drink to be consumed and systematically works through the day. Prompts are provided to aid memory of actual food or drink consumed, the composition of dishes, the cooking method, any additional items added (eg, spread on bread, type of milk in drinks, fat added to vegetables). The subject is asked to estimate portion sizes using household measures; the dietitian uses their knowledge of portion sizes to aid the subject on their recall. The recall is complete when the last food or drink consumed is detailed.
28
GRlFFITHS
Analysis of Nutritional Content Both 3-day diet diaries and 24-hour recall were then coded and analyzed blind using the commercial software package MICRODIET (Salford University 1989). The 3-day food diaries were analyzed exactly as written, no assumptions were made regarding the possible food additions (eg, spread on bread unless recorded by the patient). If the portion sizes were not stated, a medium portion using Ministry of Agriculture, Fisheries and Food (MAFF) food portion sizesl” was assumed. All analyses were completed by the same dietitian and double checked to avoid coding and data inputting errors.
Mid-Arm Circumference Mid-arm circumference (MAC) was measured using the dominant arm, with the arm hanging freely to the side and height and actual weight were recorded. These could be compared with previous measurements obtained over the preceding 6 to 12 months to ensure that dry body weight was stable.9
PCR PCR was calculated using the equation derived from detailed nitrogen balance studies in patients on continuous ambulatory peritoneal dialysis: PCR
(g/day)
= (0.261
. UA
(mmol/day))
+ 13 + TPL
(g/day),*
where UA is the total (urine plus dialysate) urea appearance and TPL is the total (urine plus dialysate) protein loss over 24 hours. PCR was normalized for dry body weight (NPCR).
Analytical Methods Plasma and dialysate concentrations of urea, creatinine, and glucose were determined on an automated discrete random access analyzer (DAX 72; Bayer Instruments, Basingstoke, UK). Urine and dialysate total protein estimations were made using standard laboratory methods.
ET
AL
Correlations between the two methods and between the dietary and urea kinetic estimates of protein intake were made using linear regression. Agreement between the two methods was analyzed by calculation of the mean and standard deviation of the differences between the observations using the method of Bland and Altman.ll
Results Patients Thirty patients, 12 men and 18 women, age 21 to 74 years, were recruited to the study. The mean period on peritoneal dialysis was 21.5 months, and they represented a broad spectrum of primary renal diseases including diabetes mellitus.
Completion of Food Diaries and 24-Hour Recall In this study, 83% patients (25 of 30) completed the 3-day food diaries. Reasons for not completing them included “forgetting to do it,” “did not have time to do it,” and absence of literacy skills; 97% (29 of 30) 24-hour recalls were obtained. One patient was unavailable for interview. The time taken to analyze the 3-day diet diary was 60 minutes compared with 30 minutes for the 24hour recall.
Assessment of Energy Intake The mean daily energy intake obtained from 3-day food diaries was 1757 kcals (7.35 MJ) compared with a mean of 1897 kcals (7.94 MJ) from 24-hour recall (Table 1). Although the Table 1. Comparison of Different Methods and Calorie Intake Method
Comparison between mean values for dietary protein and calorie intake using the two methods were made using paired Student t test, as these data were found to have a parametric distribution.
of Assessment
Protein (g/d) 24-hour recall 3-day diet diary Protein Catabolic Rate Energy intake kcals/d (MJ/d) 24-hour recall 3-day
Statistical Analysis
of Mean Values and Ranges to Measure Dietary Protein
diet diary
Mean 76.6 72.4+$ 83.4 18979 (7.94) 1757 (7.35)
Range 34.7-i 53.9 43.0-131 .o 44.3-125.1 1091-2975 (4.56-12.45) 1141-2648 (4.77-11.10)
Note. The ttest was used. *24-hour recall versus PCR, P = .19. T3-day diet diary versus PCR, P = .16. $3-day diet diary versus 24-hour recall protein, §3-day diet diary versus 24-hour recall energy,
P = .9. P = .88.
COMPARISON
Table
2. Correlation
Matrix
Variable
24-Hour Protein Intake
3-Day
Protein
NPCR
(g/kg/d)
3-Day
Energy
intake
r = P = r= P =
NPCR,
METHODS
24-Hour Energy Intake
NPCR Wgld)
.58 .0026 .58 .0009
r= .57 P = ,003
r= .78 P < .00001 Normalised
of Protein
Protein
Catabolic
Rate.
intake
The mean daily protein intake obtained from 3-day diet diaries was 72.4 g compared with 76.6 g from 24-hour recall (Table 1). As for energy intakes, there was a strong linear correlation between the two techniques (Table 2) with close agreement between paired observations (see Bland and Altman analysis, Table 3). Again, there was no evidence of systematic error. Both methods of estimating dietary protein intake were compared with that obtained from the measured PCR. The mean PCR was 83.4 g/d, approximately 10% higher than for both dietetic assessments. Using Bland and Altman analysis, the mean protein intakes from both methods were compared with PCR, these were all normalized for actual body weight. There was no significant difference between the results obtained from both methods (Table 3). Table 3. Comparison -
of Methods
Comparison Energy Protein Protein Protein
intake intake intake intake
(24-hour v3-Day) (24-hour v9Day) (24-hour vNPCR) (3-day vNPCR)
29
ASSESSMENT
This study compared two methods of dietary assessment of total calorie and protein intake in this group ofpatients receiving peritoneal dialysis. Traditionally, in the absence of a gold standard in dietary assessment, 3-day food diaries have been considered to be a suitable alternative7 and have been used in many studies, because they capture intake on different days and dietary intake is documented at the time of eating or drinking, relying less on memory. However, they rely on commitment from the patient to complete them and the patients need to be literate. Patients admit keeping them retrospectively or asking relatives/ friends to complete them thus decreasing validity. The results of this study show that the information obtained from both methods was similar with close agreement in both the mean and range of values for calorie and protein intake. As would be anticipated, there was a strong linear correlation between the methods, but also when subjected to Bland and Altman analysis, the mean differences (and their standard deviations) were small. It was also possible to exclude a systematic error using this statistical technique. One of the limitations of this study is that the 24-hour recall was performed on day 3 of the diet diary, so that the 24-hour recall may have been more accurate because the patient had previously written down this intake and would be more aware and able to remember it in better detail. Equally, the patients knew that they would be discussing their dietary intake, which may have potentially influenced the detail in their diet diary, so that it was completed in greater detail. This problem will always be encountered when patients know they are participating in a study and care needs to be taken in extrapolating findings to what can be achieved in the actual clinical situation. Bias could have been reduced by doing the 24-hour recall and the 3-day diet diary at different times; however, the study aimed to compare the
energy intake estimated from 24-hour recall was marginally greater, this did not reach of statistical significance. There was a strong and highly significant linear correlation between energy intakes estimated by the two techniques (Table 2). This was associated with a small mean difference in the daily energy intake (Table 3), with no evidence of a systematic error between the two methods of measurement (no correlation observed between the average and difference of paired measurements).
Assessment
OF DIETARY
Discussion
Intake
Abbreviation:
OF TWO
Using Bland and Altman Analysis Correlation Difference
Unit
Mean Difference
Standard Deviation
n
kcalfkgld g/kg/d glkgld g/kg/d
2.04 ,066 -.I4 -.197
6.67 .38 .39 .36
24 24 29 25
r .22 .I1 .21 -.005
Between and Mean PValue .31 .59 .27 .98
30
GRIFFITHS
protein intakes attained by both methods with the same assessment of protein catabolic rate, so the two methods of dietary assessment needed to look at the same time period. In an attempt to validate both techniques further, both dietary methods of assessing protein intake were compared with the protein catabolic rate as derived from urea kinetics and dialysate protein losses in those patients whose body weight was stable as judged by stable mid arm circumference. In both cases the dietary protein intake was less than that estimated by PCR, on average close to lo%, although this did not reach statistical significance. The reason for this discrepancy is likely to be that the method chosen to calculate PCR in this study,* which also takes into account dialysate protein losses as well as urea generation. There was no systematic effect related to patient size. Again, both dietary methods correlated with NPCR in a similar fashion. Taken altogether, these results imply that in the clinical setting both 24-hour recall and 3-day diet diaries are equally as good in estimating both protein and energy intakes when conducted by an experienced dietitian. Concerns that 24-hour recall takes too long to perform were not shown to be the case; indeed 24-hour recall required half the time of 3-day food diaries to collect, code, and input the data. In the clinical setting, the preferred method of assessment needs not only to be quick, but also to yield a good response rate from the patient. In this study, the response rate was better for 24-hour recall (97%) than 3-day diet diaries (83%). Those patients who cannot or do not complete 3-day diet diaries may be at most nutritional risk and may be missed. With 24-hour recall, the dietitian is asking the questions and therefore not relying on the patient’s literacy or motivation to complete the diet diary. However, 24-hour recall relies on the availability of an experienced dietitian. The real advantages of the technique of 24hour recall is in the face-to-face contact with the patient thus offering the opportunity to be objective in clarifying the actual food items eaten, their quantities, and how this compares to a normal intake for that individual patient. It enables the dietitian to identify other specific problems (eg, medical or social) and any change in circumstance that may effect nutritional intake which can then be highlighted to the multidisciplinary team.
ET AL
There is an opportunity for education on both diet, and other aspects such as fluid balance and compliance with phosphate binders. A face-toface contact with the patient also offers the opportunity for performing a subjective global assessment.12,13 This allows detection of any changes in physical appearance, for example signs of subcutaneous fat loss or muscle wasting, which will highlight deterioration in nutritional status and alert the multidisciplinary team that changes to treatment may be required. This is of real value in dialysis patients where levels of malnutrition are known to be high.14
Conclusion This study shows that for peritoneal dialysis patients, there is no significant difference in the estimation of protein and energy intake obtained using 24-hour recall or 3-day diet diaries. A 24-hour recall offers other advantages over 3-day diet diaries in the form of education, identifying other potential problems impacting on the care of the patient, and an opportunity for subjective global assessment. Therefore, in our experience, 24-hour dietary recall seems to be an effective method of dietary assessment and is the method of choice.
Acknowledgment The authors thank North Staffordshire Medical Institute in funding a Grant for MJB to undertake the study and Baxter Healthcare, UK, in supporting Anne Griffiths. We are also grateful statistical
to Professor
Peter
Jones,
Keele
University,
for
his
advice.
References 1. Young GA, Kopple JD, Lindholm B, et al: Nutritional assessment of continuous ambulatory peritoneal dialysis patients: An international 2. Churchill
study. AmJ Kidney Dis 12:462-471, D, Taylor DW, Keshaviah PR (The
1991 CANUSA
Peritoneal Dialysis study group): Adequacy of dialysis and nutrition in continuous peritoneal dialysis patients associated with outcome. J Am Sot Nephrol7:198-207,1996 3. Davies SJ, Russell L, Bryan J, et al: Comorbidity, urea kinetics and appetite in continuous ambulatory patients: their interrelationships and prediction Kid Dis 26:353-361,1995 4. Medlin C, Skinner ogy: A 50-year review
JD: Individual dietary intake methodolof progress. J Am Diet Assoc 88:1250-
1257,1988 5. Nelson M, Margetts methods section of dietary 6~79-83,
1993
peritoneal dialysis of survival. Am J
BM, Black investigations
AE: Checklist for the (letters). J Hum Nutr
COMPARISON 6. Fehily
AM:
Methods
of food
intake
OF TWO measurement:
uses and abuses. BNF Nutr Bullet 18:25-33. 1993 7. Bingham S: The dietary assessment ofindividuals: accuracy, Rev~ws
new techniques 57:705-743, 1987
8. Bergstrom intake, nitrogen
and
recommendations.
METHODS their
methods, Nut
dialysis. AJ,
10. Crawley Stationary
Absts
J, Fursi P, Alvestrad A, et al: Protein and energy balance and nitrogen losses in patients treated
with continuous ambulatory peritoneal 44:1048-1057,1993 9. Bennett SE, Russell GI, Williams
OF DIETARY
Kidney
Int
et al: Accurate
Office
portion 1988
sizes: London:
Her
Majesty
Statist& method for assessing of clinical management. Lancet JP, et al: What is status? J Parent Ent
Nutr 11:8-13, 1987 13. Enia G, Sicuso G, Alati G, et al: Subjective global assessment of nut&on in dialysis patients. Nephrol Dial Tramp1
Nurse Association-European ERCA) 12:137-140,1983
peritoneal nutritional
(EDTNA-
Food
(HMSO),
12. Detsky AS, Mclaughlin JR, Baker subjective global assessment of nutritional
8:1094-1098,1993 14. McCusker
Care Association
H:
11. Bland JM, Altman DG: agreement between two methods 1:307-310,1986
assessment of dry weight in haemodialysis and continuous peritoneal dialysis patients. Proc European Dialysis & Transplant Renal
31
ASSESSMENT
FX, Teehan
BP, Thorpe
KE, et al: How
dialysis is required for the maintenance state? Kidney Int 5O:S56-S61, 1996
of
much good
Objective: To conduct a comparison of two methods of dietary assessment in patients on peritoneal dialysis. Design: Comparative, cross-sectional study of two methods of dietary assessment (3-day diet diary and 24-hour recall). Data was collected simultaneously by a single experienced dietitian. Each assessment was coded and analyzed blind. Setting: Regional speciality peritoneal dialysis training unit. Patients: In this study, 30 peritoneal dialysis patients recruited prospectively and consecutively as they attended for out-patient assessment of dialysis adequacy. Age range was 22 to 77 years. Patients were excluded if unwell, younger than 18 years, or had peritonitis. Outcome measures: Total energy and protein intakes from both methods were compared. Protein intakes from both methods were compared with the protein catabolic rate generated from urea kinetics. Data obtained from both methods were compared using paired ttests, linear regression, and Bland and Altman techniques. Results: There were no significant differences in the mean daily protein (72.4 g ~76.6 g) and total energy (1757 kcals (7.35 MJ) v. 1897 kcals (7.94 MJ)) intakes determined by the two methods. Positive correlations were seen between the measurements for protein intake (r = .58, P = .0026) energy intake (r = .78, P < .OOOOl), with mean differences of .066 g/kg/d (SD .38) 2.04 kcal/kg/day (SD 6.67), respectively. For both methods there was a similar positive correlation between dietary protein intake and protein catabolic rate. A 24-hour recall was more likely to result in successful collection of data (29 of 30) than 3-day diet diaries (25 of 30) and was less time consuming. Conclusion: These two methods of determining dietary protein and energy intake do not differ significantly in the information they provide. The relative success in obtaining completed records of intake, the shorter time taken and the opportunity for patient education and assessment of other nutrition related factors has led to the adoption of the 24-hour recall method in our institution. o 1999 by the National Kidney Foundation, inc.
M
ALNUTRITION is a well documented and frequent occurrence in patients on peritoneal dialysis1 in whom it has been shown to be a predictor of poor outcome2 and to be associated with a higher incidence of mortality and morbidity.3 For these reasons it is important that there is regular dietary assessment in peritoneal dialysis patients, so that those at risk can be identified and early intervention initiated. There are many methods of dietary assessment: these include 7-day weighed intakes, photographic records, 3-day patient-kept diet diaries
0 1999 by the Natiowal Kidney Fouddioiz, 1051.2276/99/0901-0005$3.00/0
26
Inc.
using household measures, diet histories, and 24-hour dietary recall. 4,5 Each has advantages and disadvantages. In the clinical setting, the method of dietary assessment needs to be quick and accurate and yield a good response rate from the patients.6 In the absence of a gold standard method of dietary assessment, 3-day food diaries are commonly used to assess dietary intake.’ They involve a patient being given a diary, writing down a description of all food and drink consumed during a 3-day period using household measures (eg, 1 tablespoon or 1 cup, etc), instead of using weighing scales. This is less restrictive for the patient than weighing each item of food and hopefully will mean that there is greater compliance from the patient to complete the diary in detail. The potential advantages are two-fold; firstly it is thought that less time is needed by the dietitian to obtain the information and secondly that patients Jmrnal
ofRenal
Nutrition,
Vol9,
No 1 (January),
1999: pp 26-31
COMPARISON
OF TWO
METHODS
do not have to rely on their memories to complete it. However, it needs motivation and compliance from the patient, and often insufficient detail about food and drink is recorded. In addition and importantly, the patient needs to be literate. Some of the problems associated with 3-day diet diaries can be avoided by using 24-hour dietary recall. This involves a dietitian asking a patient to recall their food and drink intake within the past 24 hours. Numerous questions are asked to prompt the patient and aid memory. The dietitian can also clarify at the time of the interview the portion sizes eaten, and an accurate impression of all food and drink consumed can then be made. However, this method of dietary assessment does not take into account day-to-day variations of food and drink intake and patients with poor memories may not be able to recall their intake in enough detail to enable an accurate assessment. However, in reality, a peritoneal dialysis patient’s lifestyle is punctuated by their treatment regimen, which may, as a result, lead to less variation in day-to-day dietary intake (unpublished observations). If this were the case, then 24 hour recall may give a reasonable estimate of a dialysis patients intake. As an alternative to dietetic assessment, protein intake may be calculated using urea kinetics. Estimates of the protein catabolic rate (PCR) derived from the 24-hour urea nitrogen generation and protein losses. In carefully conducted nitrogen balance studies, this has been found to correlate with the dietary protein intake. However, this has only been shown to be the case in stable patients, where there is no urea generated from muscle breakdown.’ One of the disadvantages of this approach to assessing adequate dietary intake in peritoneal dialysis patients is its failure to assess total calorie intake. The aim of this study was to conduct a comparison of 24-hour dietary recall and 3-day diet diaries using household measures to establish if these methods are equivalent in their reliability as tools for dietary assessment in peritoneal dialysis patients. To validate these dietary assessments of protein intake both methods were also compared with PCR obtained from urea kinetics in patients whose body weight was stable as judged by a stable mid-arm circumference.9
OF DlETARY
27
ASSESSMENT
Patients and Methods Patient Population Thirty patients treated by peritoneal dialysis were recruited prospectively and consecutively as they attended for outpatient assessment of dialysis adequacy. Patients were excluded if they had an acute intercurrent illness, were younger than 18 years of age, or had peritonitis within the last month. The purpose of this selection was to obtain a stable dialysis population.
Methods of Dietary Assessment: Day Diet Diaries
Three
Before assessment, each patient completed a 3-day diet diary at home. This involved the patient filling in the diary in detail to include all food and drink consumed during the 3 days immediately preceding the assessment of dialysis adequacy. This may or may not have included a weekend day. The patients were asked to describe the food items and portion size in as much detail as possible using household measures. Food diaries also contained a short questionnaire to improve the detail of written diaries, which asked about food items that patients frequently forget to detail in food diaries hence prompting memory (eg, type of spread on bread, sugar in drinks, type and size of bread loaf). A brief written explanation on how to complete 3-day diet diaries was also included, and this was explained to them by the dietitian.
24-Hour
Dietary
Recall
A single experienced dietitian performed a 24-hour recall for the day before assessment of dialysis adequacy. This involves precise questioning of the subject of the previous days food and drink intake. The recall begins from the first food or drink to be consumed and systematically works through the day. Prompts are provided to aid memory of actual food or drink consumed, the composition of dishes, the cooking method, any additional items added (eg, spread on bread, type of milk in drinks, fat added to vegetables). The subject is asked to estimate portion sizes using household measures; the dietitian uses their knowledge of portion sizes to aid the subject on their recall. The recall is complete when the last food or drink consumed is detailed.
28
GRlFFITHS
Analysis of Nutritional Content Both 3-day diet diaries and 24-hour recall were then coded and analyzed blind using the commercial software package MICRODIET (Salford University 1989). The 3-day food diaries were analyzed exactly as written, no assumptions were made regarding the possible food additions (eg, spread on bread unless recorded by the patient). If the portion sizes were not stated, a medium portion using Ministry of Agriculture, Fisheries and Food (MAFF) food portion sizesl” was assumed. All analyses were completed by the same dietitian and double checked to avoid coding and data inputting errors.
Mid-Arm Circumference Mid-arm circumference (MAC) was measured using the dominant arm, with the arm hanging freely to the side and height and actual weight were recorded. These could be compared with previous measurements obtained over the preceding 6 to 12 months to ensure that dry body weight was stable.9
PCR PCR was calculated using the equation derived from detailed nitrogen balance studies in patients on continuous ambulatory peritoneal dialysis: PCR
(g/day)
= (0.261
. UA
(mmol/day))
+ 13 + TPL
(g/day),*
where UA is the total (urine plus dialysate) urea appearance and TPL is the total (urine plus dialysate) protein loss over 24 hours. PCR was normalized for dry body weight (NPCR).
Analytical Methods Plasma and dialysate concentrations of urea, creatinine, and glucose were determined on an automated discrete random access analyzer (DAX 72; Bayer Instruments, Basingstoke, UK). Urine and dialysate total protein estimations were made using standard laboratory methods.
ET
AL
Correlations between the two methods and between the dietary and urea kinetic estimates of protein intake were made using linear regression. Agreement between the two methods was analyzed by calculation of the mean and standard deviation of the differences between the observations using the method of Bland and Altman.ll
Results Patients Thirty patients, 12 men and 18 women, age 21 to 74 years, were recruited to the study. The mean period on peritoneal dialysis was 21.5 months, and they represented a broad spectrum of primary renal diseases including diabetes mellitus.
Completion of Food Diaries and 24-Hour Recall In this study, 83% patients (25 of 30) completed the 3-day food diaries. Reasons for not completing them included “forgetting to do it,” “did not have time to do it,” and absence of literacy skills; 97% (29 of 30) 24-hour recalls were obtained. One patient was unavailable for interview. The time taken to analyze the 3-day diet diary was 60 minutes compared with 30 minutes for the 24hour recall.
Assessment of Energy Intake The mean daily energy intake obtained from 3-day food diaries was 1757 kcals (7.35 MJ) compared with a mean of 1897 kcals (7.94 MJ) from 24-hour recall (Table 1). Although the Table 1. Comparison of Different Methods and Calorie Intake Method
Comparison between mean values for dietary protein and calorie intake using the two methods were made using paired Student t test, as these data were found to have a parametric distribution.
of Assessment
Protein (g/d) 24-hour recall 3-day diet diary Protein Catabolic Rate Energy intake kcals/d (MJ/d) 24-hour recall 3-day
Statistical Analysis
of Mean Values and Ranges to Measure Dietary Protein
diet diary
Mean 76.6 72.4+$ 83.4 18979 (7.94) 1757 (7.35)
Range 34.7-i 53.9 43.0-131 .o 44.3-125.1 1091-2975 (4.56-12.45) 1141-2648 (4.77-11.10)
Note. The ttest was used. *24-hour recall versus PCR, P = .19. T3-day diet diary versus PCR, P = .16. $3-day diet diary versus 24-hour recall protein, §3-day diet diary versus 24-hour recall energy,
P = .9. P = .88.
COMPARISON
Table
2. Correlation
Matrix
Variable
24-Hour Protein Intake
3-Day
Protein
NPCR
(g/kg/d)
3-Day
Energy
intake
r = P = r= P =
NPCR,
METHODS
24-Hour Energy Intake
NPCR Wgld)
.58 .0026 .58 .0009
r= .57 P = ,003
r= .78 P < .00001 Normalised
of Protein
Protein
Catabolic
Rate.
intake
The mean daily protein intake obtained from 3-day diet diaries was 72.4 g compared with 76.6 g from 24-hour recall (Table 1). As for energy intakes, there was a strong linear correlation between the two techniques (Table 2) with close agreement between paired observations (see Bland and Altman analysis, Table 3). Again, there was no evidence of systematic error. Both methods of estimating dietary protein intake were compared with that obtained from the measured PCR. The mean PCR was 83.4 g/d, approximately 10% higher than for both dietetic assessments. Using Bland and Altman analysis, the mean protein intakes from both methods were compared with PCR, these were all normalized for actual body weight. There was no significant difference between the results obtained from both methods (Table 3). Table 3. Comparison -
of Methods
Comparison Energy Protein Protein Protein
intake intake intake intake
(24-hour v3-Day) (24-hour v9Day) (24-hour vNPCR) (3-day vNPCR)
29
ASSESSMENT
This study compared two methods of dietary assessment of total calorie and protein intake in this group ofpatients receiving peritoneal dialysis. Traditionally, in the absence of a gold standard in dietary assessment, 3-day food diaries have been considered to be a suitable alternative7 and have been used in many studies, because they capture intake on different days and dietary intake is documented at the time of eating or drinking, relying less on memory. However, they rely on commitment from the patient to complete them and the patients need to be literate. Patients admit keeping them retrospectively or asking relatives/ friends to complete them thus decreasing validity. The results of this study show that the information obtained from both methods was similar with close agreement in both the mean and range of values for calorie and protein intake. As would be anticipated, there was a strong linear correlation between the methods, but also when subjected to Bland and Altman analysis, the mean differences (and their standard deviations) were small. It was also possible to exclude a systematic error using this statistical technique. One of the limitations of this study is that the 24-hour recall was performed on day 3 of the diet diary, so that the 24-hour recall may have been more accurate because the patient had previously written down this intake and would be more aware and able to remember it in better detail. Equally, the patients knew that they would be discussing their dietary intake, which may have potentially influenced the detail in their diet diary, so that it was completed in greater detail. This problem will always be encountered when patients know they are participating in a study and care needs to be taken in extrapolating findings to what can be achieved in the actual clinical situation. Bias could have been reduced by doing the 24-hour recall and the 3-day diet diary at different times; however, the study aimed to compare the
energy intake estimated from 24-hour recall was marginally greater, this did not reach of statistical significance. There was a strong and highly significant linear correlation between energy intakes estimated by the two techniques (Table 2). This was associated with a small mean difference in the daily energy intake (Table 3), with no evidence of a systematic error between the two methods of measurement (no correlation observed between the average and difference of paired measurements).
Assessment
OF DIETARY
Discussion
Intake
Abbreviation:
OF TWO
Using Bland and Altman Analysis Correlation Difference
Unit
Mean Difference
Standard Deviation
n
kcalfkgld g/kg/d glkgld g/kg/d
2.04 ,066 -.I4 -.197
6.67 .38 .39 .36
24 24 29 25
r .22 .I1 .21 -.005
Between and Mean PValue .31 .59 .27 .98
30
GRIFFITHS
protein intakes attained by both methods with the same assessment of protein catabolic rate, so the two methods of dietary assessment needed to look at the same time period. In an attempt to validate both techniques further, both dietary methods of assessing protein intake were compared with the protein catabolic rate as derived from urea kinetics and dialysate protein losses in those patients whose body weight was stable as judged by stable mid arm circumference. In both cases the dietary protein intake was less than that estimated by PCR, on average close to lo%, although this did not reach statistical significance. The reason for this discrepancy is likely to be that the method chosen to calculate PCR in this study,* which also takes into account dialysate protein losses as well as urea generation. There was no systematic effect related to patient size. Again, both dietary methods correlated with NPCR in a similar fashion. Taken altogether, these results imply that in the clinical setting both 24-hour recall and 3-day diet diaries are equally as good in estimating both protein and energy intakes when conducted by an experienced dietitian. Concerns that 24-hour recall takes too long to perform were not shown to be the case; indeed 24-hour recall required half the time of 3-day food diaries to collect, code, and input the data. In the clinical setting, the preferred method of assessment needs not only to be quick, but also to yield a good response rate from the patient. In this study, the response rate was better for 24-hour recall (97%) than 3-day diet diaries (83%). Those patients who cannot or do not complete 3-day diet diaries may be at most nutritional risk and may be missed. With 24-hour recall, the dietitian is asking the questions and therefore not relying on the patient’s literacy or motivation to complete the diet diary. However, 24-hour recall relies on the availability of an experienced dietitian. The real advantages of the technique of 24hour recall is in the face-to-face contact with the patient thus offering the opportunity to be objective in clarifying the actual food items eaten, their quantities, and how this compares to a normal intake for that individual patient. It enables the dietitian to identify other specific problems (eg, medical or social) and any change in circumstance that may effect nutritional intake which can then be highlighted to the multidisciplinary team.
ET AL
There is an opportunity for education on both diet, and other aspects such as fluid balance and compliance with phosphate binders. A face-toface contact with the patient also offers the opportunity for performing a subjective global assessment.12,13 This allows detection of any changes in physical appearance, for example signs of subcutaneous fat loss or muscle wasting, which will highlight deterioration in nutritional status and alert the multidisciplinary team that changes to treatment may be required. This is of real value in dialysis patients where levels of malnutrition are known to be high.14
Conclusion This study shows that for peritoneal dialysis patients, there is no significant difference in the estimation of protein and energy intake obtained using 24-hour recall or 3-day diet diaries. A 24-hour recall offers other advantages over 3-day diet diaries in the form of education, identifying other potential problems impacting on the care of the patient, and an opportunity for subjective global assessment. Therefore, in our experience, 24-hour dietary recall seems to be an effective method of dietary assessment and is the method of choice.
Acknowledgment The authors thank North Staffordshire Medical Institute in funding a Grant for MJB to undertake the study and Baxter Healthcare, UK, in supporting Anne Griffiths. We are also grateful statistical
to Professor
Peter
Jones,
Keele
University,
for
his
advice.
References 1. Young GA, Kopple JD, Lindholm B, et al: Nutritional assessment of continuous ambulatory peritoneal dialysis patients: An international 2. Churchill
study. AmJ Kidney Dis 12:462-471, D, Taylor DW, Keshaviah PR (The
1991 CANUSA
Peritoneal Dialysis study group): Adequacy of dialysis and nutrition in continuous peritoneal dialysis patients associated with outcome. J Am Sot Nephrol7:198-207,1996 3. Davies SJ, Russell L, Bryan J, et al: Comorbidity, urea kinetics and appetite in continuous ambulatory patients: their interrelationships and prediction Kid Dis 26:353-361,1995 4. Medlin C, Skinner ogy: A 50-year review
JD: Individual dietary intake methodolof progress. J Am Diet Assoc 88:1250-
1257,1988 5. Nelson M, Margetts methods section of dietary 6~79-83,
1993
peritoneal dialysis of survival. Am J
BM, Black investigations
AE: Checklist for the (letters). J Hum Nutr
COMPARISON 6. Fehily
AM:
Methods
of food
intake
OF TWO measurement:
uses and abuses. BNF Nutr Bullet 18:25-33. 1993 7. Bingham S: The dietary assessment ofindividuals: accuracy, Rev~ws
new techniques 57:705-743, 1987
8. Bergstrom intake, nitrogen
and
recommendations.
METHODS their
methods, Nut
dialysis. AJ,
10. Crawley Stationary
Absts
J, Fursi P, Alvestrad A, et al: Protein and energy balance and nitrogen losses in patients treated
with continuous ambulatory peritoneal 44:1048-1057,1993 9. Bennett SE, Russell GI, Williams
OF DIETARY
Kidney
Int
et al: Accurate
Office
portion 1988
sizes: London:
Her
Majesty
Statist& method for assessing of clinical management. Lancet JP, et al: What is status? J Parent Ent
Nutr 11:8-13, 1987 13. Enia G, Sicuso G, Alati G, et al: Subjective global assessment of nut&on in dialysis patients. Nephrol Dial Tramp1
Nurse Association-European ERCA) 12:137-140,1983
peritoneal nutritional
(EDTNA-
Food
(HMSO),
12. Detsky AS, Mclaughlin JR, Baker subjective global assessment of nutritional
8:1094-1098,1993 14. McCusker
Care Association
H:
11. Bland JM, Altman DG: agreement between two methods 1:307-310,1986
assessment of dry weight in haemodialysis and continuous peritoneal dialysis patients. Proc European Dialysis & Transplant Renal
31
ASSESSMENT
FX, Teehan
BP, Thorpe
KE, et al: How
dialysis is required for the maintenance state? Kidney Int 5O:S56-S61, 1996
of
much good