Nephrology Dialysis Transplantation near Infra-red Interactance for Nutritional Assessment of Dialysis Patients

with serum transferrin (r=+0.41, P=0.016) and cho-Background. Malnutrition is a common problem in lesterol (r=+0.39, P=0.022) and marginally with dialysis patients and may affect up to one-third of serum albumin (r=+0.29, P=0.097). patients. Near-infrared interactance (NIR) is a novel Conclusions. We conclude that NIR, which can be approach to estimate body composition and per cent performed within seconds, may serve as an objective total body fat. indicator of nutritional status in haemodialysis Methods. We used near-infrared interactance (Futrex patients. More comparative and longitudinal studies 5000B) to estimate the body composition including are needed to confirm the validity of NIR measure-body fat percentage, as well as subjective global assessments in nutritional evaluation of dialysis patients. ment (SGA), anthropometric measurements including mid-arm circumference (MAC), triceps and biceps Key words: albumin; anthropometric measurements; skinfold thickness, calculated mid-arm muscle circum-dialysis; malnutrition; near-infrared (NIR); subjective ference (MAMC), body mass index (BMI), and labor-global assessment (SGA); transferrin atory values. NIR score, SGA assessment and anthropometric parameters were measured shortly after the end of a dialysis session. NIR measurement was made by placing a FutrexB sensor on the non-access upper arm for several seconds. Serum albumin, transferrin (reflected by total iron binding capacity), Introduction and total cholesterol concentrations were performed as well. Malnutrition in haemodialysis patients is common and Results. Thirty-four patients (20 men and 14 women) may affect as many as one-third of patients [1,2]. were selected from a pool of 120 haemodialysis Assessing nutritional status is important because pro-patients. Their ages ranged from 26 to 86 years (58±14 tein–calorie malnutrition is a major risk factor for years). Time on dialysis ranged from 8 months to 19 morbidity and mortality in dialysis patients [3,4]. There years (4.5±4.6 years). NIR scores were significantly are several objective methods for assessing the nutri-different in three SGA groups: (A) well-nourished, tional status; however, all have shortcomings that 32.5±6.9%; (B) mildly to moderately malnourished, hamper their systematic clinical application [1,5]. 29.2±5.3%; and (C) severely malnourished, Anthropometric methods, which are unquestionably 23.2±10.2% (P<0.001). Pearson correlation coeffi-valuable, have inherent limitations and are impractical cients (r) between the NIR score and nutritionally for many renal units. A number of recent reports have relevant parameters were significant (P<0.001) for advocated the use of near-infrared interactance (NIR) body mass index (r=+0.81), mid-arm circumference for assessment of body compositions in different groups (r=+0.74), triceps skin fold (r=+0.54), biceps skin of patients [6–8]. NIR is a non-invasive, simple …

assess overall nutritional status of all patients.The SGA,

Subjects and methods
which was originally developed to assess nutritional status in hospitalized patients post-operatively [10], has also been The patients applied to nutritionally deprived patients in other clinical settings, including haemodialysis [2,11].The assessment is Our university-hospital-affiliated dialysis programme in based on the history and physical examination as described Staten Island, New York, currently serves over 120 patients.by Detsky et al. [10].The history consists of five criteria and We selected those patients who had never changed their focuses on weight loss in preceding 6 months, gastrointestinal modality of treatment (changed to peritoneal dialysis or symptoms (anorexia, nausea, vomiting, diarrhoea), dietary transplantation), who had required no hospitalizations in food intake, functional capacity, and co-morbidities.Each the month prior to the study, who had no signs of infections of these features are scored separately in terms of A (normal or disease activity (collagen vascular disease), and who agreed or well nourished ), B (partially abnormal or moderately to participate.Thirty-four haemodialysis patients (20 male malnourished ), or C (extremely abnormal or severely maland 14 female) agreed to enroll into the study and their data nourished ).The physical examination includes three items were obtained and analysed.Our institutional review committhat focus on loss of subcutaneous fat over the triceps and tee approved the protocol and written, informed consent was mid-axillary line of the lateral chest wall, muscle wasting in obtained from all participants.
the deltoids and quadriceps, and the presence of ankle Patients age ranged from 26 to 86 years (58±14 years).oedema and/or ascites.These features are classified as: 0= They had undergone dialysis from 8 months to 19 years normal, 1=mild, 2=moderate, and 3=severe.The data are (4.5±4.6 years).All received erythropoietin, 500-15 000 weighted and the patients are then classified in terms of three units (4706±3633 units) thrice weekly, as well as oral or major SGA scores: A=well nourished, B=moderate malintravenous iron supplementation for at least 2 months prior nutrition, or C=severe malnutrition.Details on the SGA to the study.The 'dry' body weight was the average oedemafor use in dialysis patients are available in Internet as the free weight immediately at the end of the haemodialysis appendix of our recently published manuscript [2].We sessions.
performed the SGA at two different sessions; a physician and a renal nutritionist (dietitian) performed the test independently.Differences in SGA assessment results were

Nutritional assessment
noticed on four patients or less than 15% of our study population (inter-rater reliability of 0.85) and were resolved NIR is a non-invasive, simple, and rapid method of assessing by a combined assessment consensus of the two observers.the percentage of body fat and is based on light absorption and reflection using near-infrared light emission [7].We used a commercially available NIR measuring device, Futrex Laboratory evaluation 5000A/ZLB (Futrex Inc., Galthersburg, Maryland ), which is a portable, 900 g, 12×24×5.5 cm mini-computer and is The following laboratory parameters were measured on all based on technology from the United States Department of patients immediately prior to the dialysis session: serum Agriculture, with an NIR measurement estimating range albumin, total protein, total cholesterol, serum total iron between 2.5 and 50.0%[9].The main body is connected via binding capacity (TIBC ) to estimate transferrin, serum iron, a light cable to a microphone-size light-emitting sensor.The transferrin saturation ratio (iron saturation ratio), serum NIR sensor window is to be equipped with a light shield ferritin, haemoglobin, serum creatinine, and blood urea prior to placing it on the mid upper arm to ensure that no nitrogen.Post-dialysis blood urea nitrogen of the same external light interferes with the estimation of percentage dialysis session and pre-dialysis blood urea nitrogen of the body fat.Only a few seconds are required to enter patient's following dialysis session were measured as well in order to data into the mini-computer and to obtain the NIR measurecalculate the urea reduction ratio ( URR) and protein catament while the sensor remains on the arm.In our study, bolic rate (PCR).Red blood cell indices and haemoglobin, NIR measurement was made by placing the Futrex 5000B as well as albumin, cholesterol and TIBC values (colorimetric sensor on the non-access upper arm of each dialysis patient method ) were obtained by automated methods.The haemfor several seconds.All measurements were performed atocrit was measured by centrifugation.The serum ferritin between 10 and 20 min after the termination of the dialysis value was measured by an immunoradiometric assay with session.Each patient underwent only one measurement, since polyclonal reagents.Serum TIBC concentrations were used a pilot study on 10 normal and 10 dialysis subjects showed to calculate transferrin values [12].The urea reduction ratio that the NIR scores were extremely reproducible ( less than ( URR) was obtained by calculating the percentage of intra-5% variability) as long as the same arm is used.dialytic reduction of blood urea nitrogen [2,12].URR correl-Body dry weight and skin fold measurements were perates closely with Kt/V in haemodialysis patients [4].Thus, formed between 10 and 20 min immediately after termination URR was used as the indicator of the haemodialysis efficacy of the dialysis session.Biceps skin fold (BSF ) and triceps in our study.The protein catabolic rate (PCR) was calculated skinfold ( TSF ) were measured with a conventional skinfold by equation of Gotch and Sargent based on the interdialytic caliper.Mid-arm circumference (MAC ) was measured with urea appearance rate [13].a metal tape measure.All above measurements were performed three times on the non-access arm of each dialysis patient prior to NIR measurements and the average number Statistics of the three measurements was registered as the final result.Mid-arm muscle circumference (MAMC ) was derived We used Pearson's correlation r, and Spearman rank correlation coefficient (non-parametric testing with Spearman rho) according to the following formula: MAMC=MAC− (3.1415×TSF ).Body mass index (BMI ) was calculated as for selected analysis.Two-sample student t test was used for group mean comparisons between men and women.Analysis the ratio between end dialysis body weight and the square of height.
of variance (F test) was used for group mean comparison within SGA categories.Simple and multiple regression ana-The Subjective Global Assessment (SGA) was used to lyses were carried out to obtain the regression equations.significantly different between men and women.NIR-Box and whisker plot was used to depict NIR score distribumeasured body fat percentage in women (33.2±8.4%)tion within SGA categories.Descriptive statistics and regreswas about 6% higher than that in men (27.1±6.7%).sion analyses were carried out with a statistical software Similar differences between two groups were noticed (Statistica for Windows, Release 5.1, Statsoft, Inc., Tulsa, in biceps and triceps skinfolds, both slightly higher in Oklahoma).Fiducial limits are given as mean±standard women than men.Men were significantly taller, but deviation (SD).A P value of <0.05 was accepted as statistictheir body mass index (BMI ) was only slightly greater ally significant.
than the average BMI in women.There were no gender-specific, differences in mid-arm circumference (MAC ), calculated mid-arm muscle circumference

Results
(MAMC ), urea reduction ratio ( URR) or protein catabolic rate (PCR).Similarly the serum chemistries Table 1 shows summary of data for all patients as well did not show significant differences except for serum as for each gender-specific group.Women were on albumin which was 0.3 g/dl lower in women, compared to men.average 3 years older than men.NIR scores were  Table 2 shows the clinical and laboratory data in URR (r=-0.41;P=0.017).Further correlation analyses within gender-specific subgroups revealed similar three different SGA sub-groups.Seventeen patients (50%) were well-nourished (group A), 10 patients significant correlation coefficients (P<0.05) between NIR and MAC (male r=+0.59,female r=+0.92)(29%) were mildly to moderately malnourished (group B), and seven patients (21%) were poorly nourished and MAMC (male r=+0.61,female r=+0.75).
Figures 2 and 3 depict the relationship between NIR (group C ). Figure 1 shows NIR scores according to SGA semi-quantitative levels (box and whisker plot).(y axis) and biceps skinfold and mid-arm circumference (x axis) respectively, along with their simple regression The NIR scores were significantly different in the three SGA groups: group A had 32.5±6.9%, group B had equations.
There were significant regression beta values for both serum transferrin, and number of years on dialysis.

Discussion
and years dialysed (r=-0.49,P=0.004), denoting that patient who has been on haemodialysis longer had a lower body fat percentage.Older dialysis patients We showed that NIR correlated well with the SGA and all anthropometric measurements as well as most appeared to have lower NIR measured fat percentage depicted by a near significant inverse correlation nutritionally relevant laboratory parameters in haemodialysis patients.The three SGA nutritional subgroups between NIR and age (r=-0.32,P=0.067).Among the laboratory parameters, NIR showed a statistically showed significantly different NIR measured body fat percentage values having the highest NIR scores in significant correlation coefficient with serum transferrin (r=+0.41,P=0.016) and serum cholesterol (r= well-nourished patients (SGA class A) and lowest in severely malnourished patients (SGA class C ).The +0.39, P=0.022).The relationship between NIR and serum albumin (r=+0.29;P=0.097) approached correlation was especially robust for body mass index (r=0.81)and mid-arm circumference (r=0.74).NIR signficance.NIR did not show a significant correlation with protein catabolic rate (PCR); however, there was also showed significant correlations with other anthropometric parameters including triceps and biceps skin-a significant inverse correlation between NIR and   folds and mid-arm muscle circumference.There was a patients had the longer dialysis times [2].We also found significant correlations between NIR and serum significant, inverse correlation between NIR and the years dialysed.This result corresponds to our earlier transferrin and cholersterol concentrations as well as a near-significant correlation between NIR and serum study based on subjective global assessment (SGA), in which we showed that the poorly nourished dialysis albumin.These relationships would likely have been even stronger had we examined more subjects.could serve to detect time-dependent differential changes in body composition.Kaufmann et al. [20] Although our findings are preliminary, we believe that NIR might be a promising method for quickly assessing used NIR, along with anthropometric and biochemical parameters, to investigate the impact of long-term nutritional status in dialysis patients.
NIR is based upon the principles of light absorption haemodialysis on nutritional status.Lo et al. [21] and Keshaviah et al.
[22] compared different measurements and reflection using near-infrared spectroscopy [7].When electromagnetic radiation strikes a material, the of lean body mass and included NIR in their studies as well.However, none of these investigators studied energy is reflected, adsorbed or transmitted depending on the scattering and absorption properties of the the accuracy of NIR as an independent parameter for assessing nutritional status in dialysis patients.sample.Energy scattered and reflected back out of the sample contains information about the chemical com-Our study did not show a correlation between NIR and protein catabolic rate (PCR).PCR is a parameter position of the sample [7].The use of near-infrared light interactance to determine body composition was of momentary protein intake and may not reflect the overall nutritional state.The inverse correlation first investigated by Conway et al. [14], who showed that near-infrared spectral data from the biceps of between NIR and URR was another unexpected result.
However, we did not use the average URR or average the dominant arm alone resulted in correlation coefficients equal to hydrostatic densitometry (underwater PCR of the last several months, but only the last calculated value at the time of NIR.It is possible that weighing) values.Infrared measures from the biceps have been the primary site shown to correlate best this result indicates that larger, well nourished dialysis patients receive an inadequate dialysis prescription with the standard methods [6,14].The biceps site appears to be the most representative of total body [23].Conceivably estimation of nutrition with NIR could draw attention to that fact.fat, since the combination of skin and subcutaneous fat thickness at the biceps allows optimal penetration Our study is limited by small numbers and lack of longitudinal follow-up.More elaborate methods such and interactance of the low level of infrared radiation [14].In this study, we used a commercially available as dual-energy X-ray absorptiometry (DEXA) and bioelectrical impedance (BIA) are required to further version for NIR measurement ( Futrex 5000B) which is easily available in the US and most European evaluate the validity of NIR in dialysis patients.
Nevertheless, we are encouraged by these preliminary countries at a reasonable price (currently even slightly lower than the price of the commercial versions of findings.NIR is extremely user-friendly and requires only seconds to perform.NIR-estimated body fat bioelectrical impedance analysis) and is specifically designed to estimate body fat percentage on the basis percentage may replace the more cumbersome anthropometric measurements in dialysis units, especially if of NIR measurements of the upper arm [9].
The reliability of NIR in determining body composi-its validity is corroborated by longitudinal studies tion is supported by earlier studies [6,8,15].Yasukawa composition assessment methods to evaluate their

Table 1 .
Data of all patients and their gender specific sub-groups (mean±standard deviation) NIR, near infrared; SF, skin fold; MAC, mid-arm circumference; MAMC, mid-arm muscle circumference.

Table 2 .
Patients' age, years haemodialysed, and other nutritional and laboratory data (mean±SD) in terms of SGA categories circumference; MAC, mid-arm circumference; SF, skinfold; BMI, body mass index; URR, urea reduction ratio; PCR, protein catabolic rate.
2. Kalantar-Zadeh K, Kleiner M, Dunne E et al.Total iron binding accuracy for patients with cardiac disease for the capacity-estimated transferrin concentrations in dialysis patients purpose of outcome measurement and reported that correlate with the subjective global assessment of nutrition; Am NIR presented the best standard error of estimates J Kidney Dis 1998; 31: 263-272 (3.5%) and the best correlation (r=0.84) with hydro-3.Canada-USA (CANUSA) Peritoneal Dialysis Study Group.Adequacy of dialysis and nutrition in continuous peritoneal static weighing.They elaborated that, in the future, dialysis.Association with clinical outcomes.J Am Soc Nephrol NIR measurements may be a viable technique for body 1996; 7: 198-207 composition assessment in patients with cardiac disease 4. Owen WF Jr, Lew NL, Liu Y, Lowrie EG, Lazarus JM.The although NIR consistently underestimated percent urea reduction ratio and serum albumin concentration as predictors of mortality in patients undergoing hemodialysis.N Engl body fat when compared to hydrostatic weighing in .Schoenfeld PY, Henry RR, Laird NM, Roxe DM.Assessment NIR has been applied in patients with chronic renal of nutritional status of national cooperative dialysis study popufailure in earlier studies.Soreide et al. [18] used NIR lation.Kidney Int 1983; 23 [S-13]: 80-88 to show an increase total body fat during peritoneal 6. Elia M, Parkinson SA, Diaz E. Evaluation of near-infrared interactance as a method for predicting body composition; Eur dialysis.Svarstad et al. [19] used NIR scores to meas-Brooke-Wavell K, Jones PR, Norgan NG, Hardman AE. these two group of patients and concluded that NIR 5