E Carlsson 1 *, I Bosaeus 2 and S Nordgren 1

(2002) 56, 680 686 ß 2002 Nature Publishing Group All rights reserved 0954 3007/02 $25.00 www.nature.com/ejcn ORIGINAL COMMUNICATION Body composition in patients with an ileostomy and inflammatory bowel disease: validation of bio-electric impedance spectroscopy (BIS) E Carlsson 1 *, I Bosaeus 2 and S Nordgren 1 1 Department of Surgery, University of Göteborg, Göteborg, Sweden; and 2 Department of Clinical Nutrition, University of Göteborg, Göteborg, Sweden Objective: To validate bio-electric impedance spectroscopy (BIS) by comparison with other methods for determination of body water compartments in stable subjects with an ileostomy and no or minor small bowel resection for inflammatory bowel disease (IBD). Subjects: Twenty-one subjects were included, age range 36 65 y (female=male ¼ 12=9), Crohn s disease (CD), n ¼ 14, ulcerative colitis (UC), n ¼ 6 and indeterminate colitis (IDC), n ¼ 1. Methods: Fluid compartments were assessed by the use of three independent methods: BIS, dual-energy X-ray absorptiometry (DXA) and dilution techniques (DIL); tritiated water (total body water, TBW); and bromide (extracellular water, ECW), respectively. Intra-cellular water (ICW) was calculated as TBW 7 ECW. For comparison TBW was also predicted according to an empirical formula. Differences were analysed using Bland Altman plots. Results: The mean TBW values obtained from the impedance measurement differed in the order of 7 2.21 (DIL) to 1.41 (DXA) in women and 7 2.01 (DIL) to 2.61 (DXA) in men, from the measured and derived values of total body water. Prediction of TBW gave values that were close to BIS, with a mean difference of 7 0.31 in male subjects and þ 0.51 in female subjects. Assessment of ECW revealed that the mean difference between dilution and impedance was less in women than in men (P < 0.01). Conclusion: The differences between all methods to assess fluid compartments are pronounced. To further investigate the use of the method in clinical practice for dynamic monitoring of rehydration in ileostomates with acute diarrhoea, repeated measurements together with comparison with weight fluid-balance charts are suggested. Sponsorship: The study was supported by grants from the Swedish Medical Research Council (17X-03117), Göteborgs Läkarsällskap and IB and A Lundbergs forskningsstiftelse. (2002) 56, 680 686. doi:10.1038=sj.ejcn.1601378 Keywords: body composition; inflammatory bowel disease; ileostomy; bio-electric impedance spectroscopy; dual-energy X-ray absorptiometry; tritiated water; bromide Introduction Removal of the colon may be a life-saving and curative procedure in subjects with inflammatory bowel disease *Correspondence: E Carlsson, Sahlgrenska University Hospital, Colorectal Unit, Department of Surgery, 416 85 Göteborg, Sweden. E-mail: eva.carlsson@mbox308.swipnet.se Guarantor: I Bosaeus and S Nordgren. Received 7 March 2001; revised 30 October 2001; accepted 31 October 2001 (IBD). However, this is to the cost of a permanent loss of the function of the large bowel and can be to the cost of a permanent stoma. Absence of the water-absorbing capacity of the colon implies the risk of sudden and profound changes of fluid balance, eg ileostomy diarrhoea in unspecific gastro-enteritis. Moreover, recrudescence of intestinal inflammation, as in Crohn s disease, may cause chronic increase of fluid loss through the stoma. Although several studies have shown that subjects with an ileostomy and no inflammatory activity have a normal water content according to dilution and current techniques (Christie et al, 1990; Cooper et al, 1986; Nilsson et al, 1982), the increase in water

turnover constitutes a risk of dehydration in case the ileostomy discharge is not compensated for. Assessment of dehydration in clinical practice is based mainly on anamnesis, observations regarding weight loss, physical signs and a biochemical investigation. There is presently no direct method of measurement available for practical clinical use. Research methods for the determination of total body water and assessment of fluid compartments are dominated by dilution techniques (Deurenberg et al, 1995; Gudivaka et al, 1999; Heymsfield et al, 1996; Van Marken Lichtenbelt et al, 1994). These techniques are cumbersome, require time for equilibration and involve extended laboratory facilities and calculations. For these reasons they are not suitable for monitoring the dynamic clinical course in management of dehydrated patients. The development of bio-electric impedance measurements for clinical determination of total body water represents an interesting innovation. Particularly with the use of the multiple frequency-technique (bio-electric impedance spectroscopy, BIS), allowing for a detailed analysis of water compartments, the potential clinical usefulness appears considerable (Gudivaka, et al, 1999; Van Loan & Mayclin, 1992; Van Marken Lichtenbelt et al, 1994). The apparent stability of tissue hydration also makes it possible to use body fluid investigations for estimation of other compartments, such as fat free mass (FFM) and further fat mass (Wang et al, 1999a,b). However, due to the indirect character of impedance measurements it is not known whether this technique is valid in subjects with an ileostomy, or in individuals with a disturbed fluid balance. Clinical usefulness is anticipated in ileostomates with acute disturbances of fluid balance, but also in inflammatory bowel disease (IBD) subjects with other, more longstanding, alterations of body composition. A validation was therefore considered of interest. The aim of the present study was to validate BIS by comparison with other methods for determination of body water compartments in stable subjects with an ileostomy for IBD. Subjects and methods Subjects Subjects with ulcerative colitis (UC), indeterminate colitis (IDC) or Crohn s disease (CD) were identified from the IBD clinic at Sahlgrenska University Hospital, Göteborg, Sweden. Only subjects in a clinically stable condition living in the area were asked to participate in the study. Subjects were eligible if the entire colon was removed and they had an endileostomy. Exclusion criteria included diabetes, pregnancy and ongoing inflammatory activity, as reflected by clinical history and laboratory indices (C-reactive protein (CRP), serum albumin). Subjects with a small bowel resection of more than 25 cm were also excluded. None of the patients had received steroid treatment since the last operation. Twenty-five subjects fulfilled the criteria. Four subjects (two male=two female) declined because of the individual professional situation. Twenty-one subjects (12 females and nine males) consented to participate in the study, constituting the entire population of stable ileostomates. Details regarding 681 Table 1 Clinical characteristics at entry Patient no Gender Age (y) Diagnosis Years with ileostomy Weight (kg) Height (m) BMI (kg=m 2 ) Intestinal resection (cm) 1 F 44 CD 17 71.7 1.78 22.7 25 2 F 56 CD 30 59.3 1.68 21.0 0 4 F 53 CD 7 75.3 1.60 29.4 20 6 F 43 UC 21 62.7 1.59 24.8 0 7 F 43 CD 13 72.5 1.63 27.5 25 11 F 58 CD 29 58.3 1.58 23.4 0 13 F 46 UC 2 59.4 1.66 21.6 0 15 F 56 UC 39 64.0 1.69 22.4 0 16 F 64 IDC 4 54.8 1.64 20.5 0 17 F 48 CD 24 69.4 1.61 26.8 0 18 F 47 CD 18 59.4 1.55 24.7 0 19 F 49 CD 19 82.5 1.64 30.7 0 Mean 50.6 18.6 65.8 1.64 24.6 s.d. 6.7 11.1 8.4 6.1 3.3 3 M 51 UC 25 85.5 1.89 23.9 0 5 M 36 CD 8 99.3 1.85 29.0 0 8 M 65 UC 24 83.4 1.73 27.9 0 9 M 55 UC 35 79.2 1.87 22.6 0 10 M 51 CD 27 79.6 1.90 22.1 0 12 M 42 CD 18 73.6 1.75 24.0 0 14 M 65 CD 31 91.5 1.73 30.6 0 20 M 50 CD 23 106.2 1.80 32.8 0 21 M 51 CD 28 78.9 1.81 24.1 25 Mean 51.8 24.3 86.4 1.81 26.3 s.d. 16.0 8.1 25.6 6.7 7.7

682 diagnosis, type of operations and length of the resected small intestine were collected from the medical records. For clinical status see Table 1. The patients were stable in weight and none reported nutritional problems. In fact eight of the patients were overweight, (body mass index, BMI > 25); of these three were obese (BMI > 30). Four of the women were younger than 47 y. The menstrual cycle phase was not taken into consideration for the study. Two women had oestrogen supplementation. Subjects received oral and written information on the design of the study and informed consent was obtained. The ethical committee of the Medical Faculty, University of Gothenburg approved the study protocol. Body composition analysis Subjects reported to the IBD clinic at 8.00 am after a light breakfast. All investigations were performed before noon. Clear liquids orally were permitted during the investigation. Body weight was measured to the nearest 0.1 kg with a portable platform digital scale (CL-300 BMI, Carl Lidén, Gothenburg, Sweden), with the subjects in light clothing. Height was recorded using a wall-mounted stadiometer to the nearest cm. BMI was calculated from weight and height (kg=m 2 ). Fluid compartments were assessed by use of three independent methods: BIS (experimental), dual-energy X-ray absorptiometry (DXA) and dilution techniques (reference). For comparison predicted total body water (TBW) was also calculated according to the formula presented by Bruce et al (1980). Bio-electrical impedance spectroscopy. A commercially available instrument (Xitron Hydra 4200, Xitron Technologies, San Diego, USA) was used. The impedance technique is based on the observation that the electric conductivity of the human body is dependent on the frequency of an alternating current, and that resistivity of intra- and extra-cellullar water are different for different frequencies. The Xitron Hydra 4200 uses 50 frequencies between 5 khz and 1 MHz. At the highest frequencies, the current is able to overcome the capacitance of cell membranes, and penetrate throughout virtually all the body water pool, whereas at lower frequencies, the current passes only through extracellular water. In practice water volumes are calculated by inserting patient variables into an equation using constants and a computer program supplied by the manufacturer. For a detailed description see reviews by De Lorenzo et al (1997) and Chumlea and Guo (1997). Subjects were resting in the supine position during 10 min prior to assessment. Room temperature was 23 C and humidity varied between 20 and 30%. Electrodes were placed according to the manufacturer s manual: one pair of electrodes on the dorsal surface of the hand=wrist and the other on the foot=ankle on the right side of the body. Measurements were performed twice; the mean value was used. The precision error was reported to be < 2% (Ellis & Wong, 1998). Values for ECW and ICW compartments were obtained using the instrument s software option for water compartment analyses after entering of gender, height and weight. TBW was calculated as the sum: TBW (BIS) =ICW (BIS) + ECW (BIS) Dual-energy X-ray absorptiometry. DXA was performed using a LUNAR-DPX-IQ scanner (Scanexport Medical, Helsingborg, Sweden). The system used a constant-potential X- ray source and a K-edge filter to achieve a congruent beam of stable dual-energy radiation. Whole body scans were performed at the scan speed suggested by the system for each subject. Body fat, FFM, total bone mineral content and density were analysed using the software, version 4.7 C. TBW values were derived from the DXA-analysis. For the present study, hydration of the FFM was assumed to be constant: (Pace & Rathburn, 1945; Wang et al, 1999b): TBW (DXA) =0.732 FFM Dilution techniques (DIL). Tritiated water and bromide were used for assessment of TBW (DIL) and ECW (DIL) respectively. The bladder and stoma-bag were emptied, and a urine sample was collected for baseline analyses. Each subject consumed a mixture of tritiated water and 45 ml of a 5% solution of NaBr followed by a further 200 ml of tap water to wash out the cup. TBW (DIL) was determined by assessment of isotope dilution of tritiated water (THO; Bruce et al, 1980). Subjects were given 100 mci (3.7 MBq) of THO, added to the bromide solution. During equilibration (3 h) the stoma output was collected for analysis of tritiated water in 12 of the 21 patients (in nine the stomal discharge was too viscous for analysis). In all stoma samples the specific activity of tritiated water in the stoma content was less than or equal to the plasma value, in concert with an earlier paper from our institution (Nilsson et al, 1982). After equilibration a blood sample of 10 ml was drawn from a cubital vein. The activity of plasma water, obtained by sublimation, was counted in a liquid scintillation counter. The dilution technique using tritiated water over-estimates TBW by 4 5% (Schoeller, 1992). The tritium-dilution space was therefore divided by 1.05, correcting for the exchange of tritium label with nonaqueous hydrogen in body solids (Schoeller, 1992). The coefficient of variation of a single TBW determination was 3.2% (Bruce et al, 1980). Samples from two patients (one man, one woman) were omitted for technical reasons. ECW (DIL) was determined as bromide space, modified after Miller et al, (1989). After 3 h of equilibration, a blood sample was drawn, serum was separated and stored at 7 20 C until analysis. Samples were deproteinised using methanol and after subsequent centrifugation, bromide concentration was determined by high performance ion chromatography (HPIC, Dionex ion-chromatographic

system 4500i, Sunnyvale, CA, USA, equipped with a PAX-100 guard column, PAX-100 analytical column, a AMMS-II micro membrane suppressor and a conductivity detector). Data were computed using Borwin chromatography software (version 1.20, Le Fontanil, France). Samples were eluated with a gradient mixture of water, 0.2 M NaOH and 50=50 isopropanol. Standards were prepared from commercially available NaBr (Merck GmbH, Darmstadt, Germany). All chemicals were of analytical grade (p.a.), the water was type 1 reagent grade with a specific resistance of 17.8 MO=cm or greater. Samples were analysed in duplicate. Coefficient of variation in duplicate samples was 1.8%. Bromide space as a measure of extracellular volume was defined as: ECW ðdilþ ¼ ðbr given dose=br concentration in serumþ0:90 0:95 The correction factor for non-extracellular distribution of Br ¼ 0.90, and correction for the Donnan equilibrium ¼ 0.95 (Miller et al, 1989). Prediction of TBW. Predicted TBW (PRE) was calculated in all subjects according to the modified formula derived by Bruce et al (1980). This formula uses body weight and height, sex and age and gives an estimate of TBW, r (male) ¼ 0.755, r (female) ¼ 0.773. The original equation was divided by 1.05 (in accordance with what was stated above concerning dilution techniques). The formulas are: Male: TBW ðpreþ ¼ð0:40 BW ðkgþþ0:023 BH ðcmþ 0:056 Age ðyþþ12:1þ=1:05 Female: TBW ðpreþ ¼ð0:24 BW ðkgþþ0:20 BH ðcmþ 0:03 Age ðyþ 13:9Þ=1:05 Statistical analysis All statistical analyses were made using the SPSS 8.0 for Windows package (SPSS Inc., Chicago, IL, USA). Descriptive statistics were reported as mean and s.d. Means were compared using paired and unpaired t-tests. Bland Altman plots (Bland & Altman, 1986) were used to assess the degree of interchangeability between impedance, DXA and dilution methods. A line of regression was included in the figures, and Pearsons s correlation coefficient was calculated. The limits of agreement between the two methods were defined as mean difference 2 s.d. (Bland & Altman, 1986). A P- value < 0.05 was considered significant. Results Twenty-one subjects participated; 14 had CD, six UC and in one the diagnosis was IDC. Age ranged between 36 and 65 y; the mean (s.d.) age was 51 ( 7.8) y. BMI was 25.3 ( 3.6) kg=m 2. Time elapsed since the ileostomy operation was 21 ( 10) y. Table 1 shows individual clinical data for patients at entry into the study. In Table 2 the mean ( s.d.) of measurements of fluid compartments and FFM are denoted with reference to the three methods. It was observed that the mean TBW values obtained from BIS differed in the order of 7 2.21 (DIL) to þ 1.41 (DXA) in women and 7 2.01 (DIL) to þ 2.61 (DXA) in men, from the measured and derived values of TBW. DXA returned lower values and the dilution techniques returned higher values when compared to BIS. This observation was concordant in male and female subjects. A significant correlation was shown between TBW (DIL) and TBW (BIS), r ¼ 0.92, P < 0.001, and also between TBW (DXA) and TBW (BIS) r ¼ 7 0.94, P < 0.01. TBW (PRE) gave values that were close to BIS, with a mean difference of 7 0.31 in male and þ 0.51 in female subjects. The differences between the reference methods (DXA and DIL) and experimental (BIS) were analysed using Bland Altman plots, TBW (Figures 1 and 2) and ECW (Figure 3). Mean differences and lines of regression were plotted. Lines corresponding to mean 2 s.d. were also plotted in the diagrams. Detailed analysis of differences in assessment of ECW using bromide dilution and BIS revealed that the mean difference between the two techniques was significantly less in women than the corresponding difference in men, P < 0.01. A significant negative correlation between the difference TBW (DXA) 7 TBW (BIS) and BMI was observed, r ¼ 7 0.77 P < 0.01. No correlation was observed for the difference TBW (DIL) 7 TBW (BIS) and BMI, r ¼ 0.001, NS. Discussion Rapid shifts in hydration constitute a significant problem in subjects with an ileostomy. Assessment of dehydration in clinical practice is based mainly on the clinical history of recent weight-loss, although this information is frequently hampered by an inability to recall weight. Present day methods, encompassing mainly dilution techniques, are technically demanding and time-consuming. A simple and Table 2 Measurements of body composition n TBW ICW ECW FFM Female DXA 12 29.7 2.8 40.6 3.9 Impedance 12 31.1 4.3 17.4 2.9 13.7 1.5 42.6 5.9 Dilution 11 33.3 5.1 20.4 3.8 13.0 1.9 45.6 7.0 Predicted (Bruce et al, 1980) 12 31.6 2.6 Male DXA 9 43.7 3.5 59.9 4.8 Impedance 9 46.3 4.8 26.5 3.4 19.8 1.9 63.4 6.5 Dilution 8 48.3 6.5 31.2 5.4 17.0 2.1 66.1 8.9 Predicted (Bruce et al, 1980) 9 46.0 4.3 683

684 Figure 1 Bland Altman plot showing the differences between TBW (DIL) and TBW (BIS) (for explanation see text) vs mean TBW (DIL and BIS). A solid line represents mean difference. Lines representing mean 2 s.d. are horizontal and dashed. The dashed line crossing the mean is a line of regression. Females (circle); males (square). Figure 2 Bland Altman plot showing the difference between TBW (DXA) and TBW (BIS) versus mean (DXA and BIS) (for explanation see text). Lines and symbols are the same as in Figure 1. Figure 3 Bland Altman plot showing the difference between ECW (DIL) and ECW (BIS) vs mean ECW (DIL and BIS, for explanation see text). Lines and symbols are the same as in Figure 1. Note the consistent difference between men and women. accurate method for assessment of fluid status that can be used in clinical practice is therefore needed. Determination of body water with BIS has many practical advantages. The equipment is relatively inexpensive and requires limited operator training and maintenance. Measurements can be done bedside and repeatedly, as frequently as needed, and results are immediately available. The method is quick and non-invasive. Subjects do not have to drink unpleasant test solutions and to have blood samples taken. There is no need for equilibration. However the accuracy of the method in different clinical settings is as yet unknown. Reduced accuracy is described in obese subjects (Cox-Reijven & Soethers, 2000; Deurenberg, 1996) and according to Cox-Reijven and Soehters the accuracy can only be improved if the constants in the Xitron model are adjusted for the degree of overweight. They also suggest that, before all possible confounding factors are identified, BIS should be used with caution in clinical practice. A difference was observed between TBW (BIS) and TBW (DIL) in a validation of the impedance-method in a group of 469 subjects (Ellis & Wong, 1998). In that study, impedancederived values were consistently lower than those obtained by the dilution method, which is in agreement with the present study. However, further comparison is limited by the fact that the patient material was dominated by a multiethnic and pediatric group of subjects. When the impedance instrument was recalibrated using constants from values obtained using the DXA-method Bland Altman requirements (Bland & Altman, 1986) for interchangeability between methods were met. The s.d. values for the mean differences between the dilution methods and BIS methods, however, remained pronounced both in males and females.

Royall and co-workers, on the other hand, noted that impedance over-estimates TBW as recorded using dilution of H 2 18 O (Royall et al, 1994). A consistent observation in the present study was that DXA values for TBW were lower than BIS, which is in contrast to Ellis et al (1999); the magnitude of the difference, again, is similar. It was also observed that TBW (DXA) 7 TBW (BIS) demonstrated a negative regression with BMI which was not observed in TBW (DIL) 7 TBW(BIS). This limitation in the use of DXA as a reference method for deriving TBW in obese has not received attention previously. However DXA may have limitations in body composition analysis in the very obese (Jebb & Elia, 1993). The advantage in using BIS is claimed to be the possibility to calculate ECW and ICW separately. The clinical usefulness is undisputed, eg in acute situations of dehydration (Dewit et al, 1997) and with or without ongoing inflammation. The variation between methods was wide in the present study, which reduces the ability to make reliable individual measurements. A similar conclusion was reached by Gudivaka et al (1999), and Cox-Reijven and Soethers (2000) suggest that the accuracy is not high enough for routine clinical use. However, as stated by Dewit et al, although the use of BIS for precise prediction volumes may be questioned, its capacity for frequent serial use in longitudinal studies where repeated use of tracer dilution or other techniques is not possible commends its wider application (Dewit et al, 1997). Ellis also reported significant differences between male and female subjects regarding TBW (BIS) and TBW (DXA, Ellis et al, 1999; Ellis & Wong, 1998). This is in parallel with our findings regarding ECW (DIL). Using the bromide-dilution technique we observed a statistically significant difference between men and women. Gender differences are frequently found in impedance models (Gudivaka et al, 1999). Gudivaka et al found that most ECW and ICW models required gender-specific equations, which was unexpected. There may be a systematic difference in the relative contribution of the limbs and trunk in men and women, and this leads to the gender difference in regression lines. They also discussed whether it could be due to difference in the conductivity of the body fluids, but this was not supported by data. Another reason could be that there might be a gender difference in the bromide ICW penetration or mixing time that resulted in an artifact in the criterion method. They found the gender effect perplexing and concluded that more research is needed (Gudivaka et al, 1999). Durenberg et al (1995) found that the prediction of ECW from impedance index alone was improved when body weight and age were added as independent variables and the prediction of TBW could also be improved by the additional factors body weight and gender. Gallagher et al, in their study including 269 women and 204 men, found that age, body density and gender all contributed significantly to individual differences in BCM. They also found that the gender effect observed in regression models was mainly due to variation in extracellular fluid (Gallagher et al, 1996). The present study was designed to analyse the validity of BIS in a sample of subjects with an ileostomy, by comparing measurements with results obtained using reference techniques. BIS values consistently fell between values from dilution methods and DXA. Ileostomy subjects are included in a few previous reports (Christie et al, 1990; Cooper et al, 1986; Nilsson et al, 1982), but have not attracted much interest before. The IBD-population raises specific problems in the analysis of body composition: an increased susceptibility to dehydration was mentioned above. Moreover the presence of inflammation and also pharmacological treatment of inflammation may cause shifts of fluids and changes in body composition that make use of the BIS technique unreliable. The ECW=ICW ratio in patients with a long history of Crohn s disease showed a slightly increased value in a study by Geerling et al (1999). In the present study all subjects had an ileostomy, whereas only two out of 17 had an ileostomy in the quoted study. These circumstances necessitate a careful validation of BIS as the measurement is indirectly based on electric properties of the various tissues of the human body, the effects on resisitivity cannot be predicted. It was noted that in the present study subjects were not clinically dehydrated, had no ongoing inflammation and did not take any anti-inflammatory drugs, including steroids. The small sample in this study does not allow for fully reliable statistical conclusions. The number of subjects was similar to most previous studies on stable ileostomates (Cooper et al, 1986; Hill et al, 1975; Nilsson et al, 1982). It has been claimed that subjects with an ileostomy are chronically dehydrated (Hill et al, 1975), although the topic is controversial. Ileostomy subjects have been investigated for dehydration without having recorded any significant differences (Cooper et al, 1986; Nilsson et al, 1982). Although the number of patients under study is small, the previous conclusion that ileostomy subjects have a normal body composition and the present observations on the validity of BIS make it reasonable to suggest that BIS can be used in these patients, at least under assumed normal hydration. In the clinical situation it seems reasonable from the present observations that expected TBW (under steady-state conditions) may be calculated from the equation by Bruce et al (1980) and restitution during fluid therapy may be monitored using BIS. The differences between all methods to assess fluid compartments are pronounced. To further investigate the use of the method in clinical practice for dynamic monitoring of rehydration in ileostomates with acute diarrhoea repeated measurements during the course together with comparison with weight fluid-balance charts are suggested. This will be addressed in a future study. Acknowledgements This investigation was supported by grants from the Swedish Medical Research Council (17X-03117), Göteborgs 685

686 Läkarsällskap, and IB och A Lundbergs Forskningsstiftelse. Mr NG Carlsson, who is gratefully acknowledged, performed the bromide analyses. References Bland J & Altman D (1986): Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 8, 307 310. Bruce Å, Andersson M, Arvidsson B & Isaksson B (1980): Body composition. Prediction of normal body potassium, body water and body fat in adults on the basis of body height, body weight and age. Scand. J. Clin. Lab. Invest. 40, 461 473. Christie P, Knight G & Hill G (1990): Metabolism of body water and electrolytes after surgery for ulcerative colitis: conventional ileostomy vs. J-pouch. Br. J. Surg. 77, 149 151. Chumlea C & Guo S. (1997): Bioelectric impedance: a history, research issues, and recent consensus. In Emerging Technologies for Nutrition Research, ed. S Carlson-Newberry & R Costello, pp 169 192. Washington, DC: National Academy Press. Cooper J, Laughland A, Gunning E, Burkinshaw L & Williams N (1986): Body composition in ileostomy patients with and without ileal resection. Gut 27, 680 685. Cox-Reijven P & Soethers P (2000): Validation of bio-impedance spectroscopy: effects of degree of obesity and ways of calculating volumes from measured resistance values. Int. J. Obes. Relat. Metab. Disord. 24, 271 280. De Lorenzo A, Andreoli A, Matthie J & Withers P (1997): Predicting body cell mass with bioimpedance by using theoretical methods: a technological review. J. Appl. Physiol. 82, 1542 1558. Deurenberg P (1996): Limitations of the bioelectrical impedance method for the assessment of body fat in severe obesity. Am. J. Clin. Nutr. 64, 449S 452S. Deurenberg P, Tagliabue A & Shouten F (1995): Multi-frequency impedance for the prediction of extracellular water and total body water. Br. J. Nutr. 73, 349 358. Dewit O, Ward L, Middleton SJ, Watson C, Friend PJ & Elia M (1997): Multiple frequency bioimpedance: a bed-side technique for assessment of fluid shift patterns in a patient with severe dehydration. Clin. Nutr. 16, 189 192. Ellis K & Wong W (1998): Human hydrometry: comparison of multifrequency bioelectrical impedance with 2 H 2 O and bromine dilution. J. Appl. Physiol. 85, 1056 1062. Ellis K, Shypailo R & Wong W (1999): Measurement of body water by multifrequency bioelectrical impedance spectroscopy in a multiethnic pediatric population. Am. J. Clin. Nutr. 70, 847 853. Gallagher D, Visser M, Wang Z, Harris T, Pierson R & Heymsfield S (1996): Metabolically active component of fat-free body mass: influences of age, adiposity, and gender. Metabolism 45, 992 997. Geerling B, van Marken Lichtenbelt W, Stockbrugger R & Brummer R-JM (1999): Gender specific alterations of body composition in patients with inflammatory bowel disease compared with controls. Eur. J. Clin. Nutr. 53, 479 485. Gudivaka R, Schoeller D, Kushner R & Bolt M (1999): Single- and multi-frequency models for bioeletrical impedance analysis of body water compartments. J. Appl. Physiol. 87, 1087 1096. Heymsfield S, Wang Z, Visser M, Gallagher D & Pierson R Jr (1996): Techniques used in the measurement of body composition: an overview with emphasis on bioelectrical impedance analysis. Am. J. Clin. Nutr. 64, 478 484. Hill G, Goligher J, Smith A & Mair W (1975): Long term changes in total body water, total exchangeable sodium and total body potassium before and after ileostomy. Br. J. Surg. 62, 524 527. Jebb S & Elia M (1993): Techniques for the measurement of body composition: a practical guide. Int. J. Obes. Relat. Metab. Disord. 17, 611 621. Miller M, Cosgriff J & Forbes G (1989): Bromide space determination using anion-exchange chromatography for measurement of bromide. Am. J. Clin. Nutr. 50, 168 171. Nilsson L, Andersson H, Bosaeus I & Myrvold H (1982): Total body water and total body potassium in patients with continent ileostomies. Gut 23, 589 593. Pace I & Rathburn E (1945): Studies on body composition. III The body water and chemically combined nitrogen content in relation to fat content. J. Biol. Chem. 158, 685 691. Royall D, Greenberg G, Allard J, Baker J, Harrison J & Jeejeebhoy K (1994): Critical assessment of body-composition measurements in malnourished subjects with Crohn s disease: the role of bioelectric impedance analysis. Am. J. Clin. Nutr. 59, 325 330. Schoeller D (1992): Isotope dilution methods. In Obesity, ed. B Björntorp & B Brodoff, pp 80 88. New York: Lippingcott. Van Loan MD & Mayclin PL (1992): Use of multi-frequency bioelectrical impedance analysis for the estimation of extracellular fluid. Eur. J. Clin. Nutr. 46, 117 124. Van Marken Lichtenbelt WD, Westerterp KR, Wouters L & Luijendijk SC (1994): Validation of bioelectrical-impedance measurements as a method to estimate body-water compartments. Am. J. Clin. Nutr. 60, 159 166. Wang Z, Deurenberg P, Wang W, Pietrobelli A, Baumgartner RN & Heymsfield S (1999a): Hydration of fat-free body mass: new physiological modeling approach. Am. J. Physiol. 276(Endocrinol. Metab 39), E995 E1003. Wang Z, Deurenberg P, Wang W, Pietrobelli A, Baumgartner RN & Heymsfield S (1999b): Hydration of fat-free body mass: review and critique of a classic body-composition constant 1 3. Am. J. Clin. Nutr. 69, 833 841.