European Journal of Clinical Nutrition (1997) 51, 810±814 ß 1997 Stockton Press. All rights reserved 0954±3007/97 $12.00 Whole body and regional body composition analysis by dual-energy X-ray absorptiometry in cirrhotic patients

O Riggio1, A Andreoli2, F Diana1, P Fiore1, P Meddi1, R Lionetti1, F Montagnese1, M Merli1, L Capocaccia1 and A De Lorenzo2

1II Gastroenterologia, UniversitaÁ, La Sapienza and 2Fisiologia Umana, UniversitaÁ di Tor Vergata, Rome, Italy

Objectives: To compare whole body and regional (arms, legs and trunk) fat mass, fat-free mineral-free mass, mineral content and bone mineral density, measured by DXA, in cirrhotic patients and age, sex and BMI matched healthy volunteers. Design: Cross-sectional study. Setting: Two medical research institutions. Subjects: Twenty-two non ascitic cirrhotic patients and 16 age, sex and BMI matched healthy volunteers. Interventions: The Lunar DPX whole-body X-ray densitometer with Lunar software version 3.6z (Lunar Radiation Corp., Madison WI, USA) was used. Regional analysis was performed on the arms, legs, trunk and head. Results: Compared to controls, cirrhotic patients showed a signi®cant reduction in percentage body fat. When differentiated by gender, however, the reduction in percentage body fat was evident in female cirrhotics only, particularly in the trunk. In male cirrhotic patients fat-free mineral-free mass was reduced in absolute terms in the whole body and the limbs. For both genders and in each body segment bone mineral content and density were reduced in cirrhotics compared to controls. In cirrhotic patients bone mineral density was signi®cantly correlated to both fat-free, mineral-free mass (r ˆ 0.85; P < 0.001) and to the Physical Activity Index (r ˆ 0.52; P < 0.01). Conclusions: Two different patterns of soft tissue loss may be found in cirrhotic patients: in women lean tissue is maintained while fat stores are reduced, as in early starvation; in men lean tissue is reduced, as seen under conditions of stress. Moreover, factors in¯uencing , such as nutritional depletion and physical inactivity, may contribute to the reduction of bone density frequently observed in cirrhotic patients. Sponsorship: This work has been funded by the University of Rome, `La Sapienza' research grant MURST 60%, 1995. Descriptors: body composition; dual-energy X-ray absorptiometry; cirrhosis

Introduction hydratation is variable (Katch et al, 1986; Jebb & Elia, 1991; De Lorenzo et al, 1991), as is the case with cirrhotic The measurement of body composition in cirrhotic patients patients (McCullough et al, 1991; Olroyd et al, 1993; is of particular interest, not only because it can provide Prijatmoko et al, 1993; Guglielmi et al, 1991). The estima- additional and more precise information on their nutritional tion of body composition from Total Body Potassium status, but also because the estimation of metabolically (TBK) or Total Body Water (TBW) is based on assump- active body compartments by means of body composition tions made regarding healthy individuals and has never analysis is essential in standardizing physiological pro- been validated in cirrhotic patients in whom TBK may be cesses, such as energy expenditure and protein turnover. reduced (Podolsky et al, 1973) and fractional hydratation A comparison of these processes between cirrhotics and of body lean and cell mass may be increased (McCullough healthy individuals may help explain the mechanism by et al, 1991; Olroyd et al, 1993; Prijatmoko et al, 1993). which malnutrition arises in these patients. Finally, the application of in vivo neutral activation analysis Although several methods have been applied (Madden (IVNAA) can only be performed in a few specialized & Morgan, 1994; Muller et al, 1992; McCullough et al, centers and thus in a limited number of patients. 1991; Crawford et al, 1994; Olroyd et al, 1993; Prijatmoko Dual X-ray absorptiometry is a safe, convenient and et al, 1993), an analysis of body composition in cirrhotic non invasive technique for assessing body composition patients is not easy to perform. Simple bed-side methods, (Mazess et al, 1990). By applying this method Fat Mass skinfold anthropometry for example, while useful for (FM), Bone Mineral Content (BMC), namely mass in epidemiological purposes (Italian Multicentre Cooperative grams of the inorganic calci®ed content of bone, Bone Project on nutrition in liver cirrhosis, 1994; Merli et al, Mineral Density (BMD), namely bone mass per unit of 1996), are not always accurate (Frost & Corish, 1989). projected bone area in g=cm2) and Fat-Free Mineral- Bioelectrical impedance analysis may not provide valid Free Mass (FFMFM) can be estimated, and whole body data on body composition in patients whose state of and regional analysis (arms, legs and trunk) may be performed. This study aimed to determine the above three body Correspondence: Dr O Riggio, Via Costantino MAES 68, 00162 Roma, Italy. compartments in non ascitic cirrhotic patients and age, Received 30 December 1996; revised 8 July 1997; accepted 22 July 1997 gender and body size-matched controls. Body composition analysis in cirrhotic patients O Riggio et al 811 Table 1 Etiology and functional severity of liver injury in the 22 cirrhotic patients

Etiology Patients

HCV 10 Alcoholic 4 Cryptogenetic 2 HCV ‡ alcoholic 2 HBV ‡ HDV 2 HBV ‡ alcoholic 1 PBC 1 Child±PUGH A 9 B 12 C 1 Bilirubin (mg/dl) 2.8 Æ 3.8 Prothrombin activity (%) 66.7 Æ 14.6 Albumin (g/dl) 3.4 Æ 0.38

HCV ˆ Hepatitis C virus; HBV ˆ Hepatitis B virus; HDV ˆ Hepatitis Delta virus; PBC ˆ Primary Biliary Cirrhosis. Figure 1 Correlation between whole body mass by DXA and body weight by gravimetry. Table 2 Comparison between cirrhotic patients and matched healthy volunteers between cirrhotics and matched controls is reported in Cirrhotics Controls Table 2. Sex (M/F) 12/10 7/9 DXA measurements were made on subjects in the Age 61.2 Æ 7 60.5 Æ 10.6 NS postabsorptive state immediately after emptying the blad- Weight (kg) 71 Æ 15 75 Æ 11 NS der and with no ¯uid intake in the previous 4 hrs. The Height (cm) 165 Æ 9 166 Æ 9 NS BMI (kg/m2) 26 Æ 5 27 Æ 2 NS Lunar DPX whole-body X-ray densitometer with Lunar software version 3.6z (Lunar Radiation Corp., Madison WI, Mean Æ SD. USA) was used. This instrument used a rectilinear scanner, running at medium speed, to detect density differences through the subject, lying undressed on a scan table. Regional analysis was performed on the arms, legs, trunk Patients and methods and head. This analysis provides mass in grams of bone Twenty-two cirrhotic patients hospitalised for diagnostic mineral content, fat, lean (fat-free, mineral-free mass) and and=or therapeutic reasons were included in the study. the sum of the body tissues for segments and whole body. Patient's characteristics are reported in Table 1. Aetiology The determination of whole body mass by DXA and body of liver disease was based on histological, clinical and weight by gravimetry were compared (Figure 1). laboratory variables. The severity of cirrhosis was assessed Each patient ®lled in a questionnaire regarding their using the Child-Pugh grading system, performed in each usual daily routine. Questions were asked about how many patient on entry. Ten of these patients had clinical and hours are spent sleeping, awake but in repose or in physical ultrasonographical evidence of ¯uid retention at entry. activity (housework, personal care, outdoor pursuits, etc.) Fluid retention was treated with diuretics or paracenthesis (Wilson et al, 1986). An index (in percent) was calculated and body composition measurements were made when by simply dividing the number of hours devoted to physical ascites (at ultrasonography) or oedema was no longer activity by 24. detectable. Sixteen healthy volunteers of similar age and Data from cirrhotics and controls were compared using were used as controls. The comparison the Student t-test. All results are espressed as mean Æ SD

Table 3 Regional analysis of body composition in cirrhotics and controls (MeanÆSD)

BMC BMD FM FF-MFM Total Fat (kg) (g=cm2) (kg) (kg) (kg) ( % )

Cirrhotics 0.36 Æ 0.10 0.86 Æ 0.12 3.33 Æ 2.26 5.00 Æ 1.49 8.70 Æ 3.39 35.90 Æ 11.80 Arms Controls 0.39 Æ 0.14 0.93 Æ 0.10 4.03 Æ 1.56 5.41 Æ 1.86 9.83 Æ 2.57 41.50 Æ 11.50 t-test ns ns ns ns ns ns

Cirrhotics 0.61 Æ 0.16 0.82 Æ 0.01 10.16 Æ 4.5 24.41 Æ 4.72 35.18 Æ 7.54 20.12 Æ 8.40 Trunk Controls 0.79 Æ 0.20 0.94 Æ 0.01 13.32 Æ 3.3 22.94 Æ 5.12 37.05 Æ 6.25 36.16 Æ 7.40 t-test P < 0.005 P < 0.0005 P < 0.05 ns ns P < 0.005

Cirrhotics 0.90 Æ 0.24 1.15 Æ 0.17 6.66 Æ 2.71 13.85 Æ 2.88 21.41 Æ 4.64 30.51 Æ 8.90 Legs Controls 1.04 Æ 0.29 1.26 Æ 0.15 7.05 Æ 2.16 14.87 Æ 4.30 22.97 Æ 4.58 31.40 Æ 9.30 t-test ns ns ns ns ns ns

Cirrhotics 2.32 Æ 0.50 1.08 Æ 0.10 21.15 Æ 9.08 46.29 Æ 8.80 69.77 Æ 15.12 29.49 Æ 8.20 Total body Controls 2.71 Æ 0.60 1.18 Æ 0.09 25.39 Æ 5.07 46.22 Æ 11.10 74.33 Æ 11.20 34.77 Æ 7.62 t-test P < 0.05 P < 0.01 ns ns ns P < 0.05

BMC ˆ Bone Mineral Content; BMD ˆ Bone Mineral Density; FM ˆ Fat Mass; FF-MFM ˆ Fat Free Mineral-Free Mass. Body composition analysis in cirrhotic patients O Riggio et al 812 5), BMC and BMD were signi®cantly reduced in both genders and in each body segment compared with respec- tive controls, while percent body fat was signi®cantly reduced in female cirrhotics only, particularly in the trunk. In male cirrhotic patients fat-free, mineral-free mass was lower in absolute terms but not in percentage in the whole body or in the limbs (Table 4). In the cirrhotic group the BMD was signi®cantly corre- lated to both body weight (r ˆ 0.65; P < 0.001) and fat-free, mineral-free mass (r ˆ 0.85; P < 0.001), but not to body fat (r ˆ 0.19; ns) (Figure 2). BMD was also correlated to the physical activity index (r ˆ 0.52; P < 0.01). The mean physical activity index of cirrhotic patients was 18 Æ 10%.

Discussion DXA has been introduced as a precise, safe and relatively inexpensive tool to measure body composition in humans. DXA is associated with a low radiation exposure (Lang et al, 1991) and provides bone mineral as well as fat and lean tissue mass, and thus body composition may be estimated according to the three-component model (Lukasky, 1993). DXA has been validated in animal studies where carcass analysis is possible (Svendsen et al, 1993) and in humans by comparing other methods for estimating body composi- tion, such as under-water weighing (Johansson et al, 1993), skinfold thickness, bioelectrical impedance and TBW (Pritchard et al, 1993), TBK and TBN (Aloia et al, 1995). Both the precision and the agreement of DXA with other methods are considered excellent for bone mineral content, and very good for fat mass. On the other hand, DXA tends to underestimate lean mass, especially at highest values (Aloia et al, 1995). In cirrhotic patients DXA has been compared to TBK, BIA and skinfold anthropometry (Bramley et al, 1993) and to a multicom- partmental model based on IVNAA and TBW measure- ments (Prijatmoko et al, 1993). Again a good agreement for Figure 2 Correlation between whole body BMD and Fat-Free Mineral- fat mass estimation was found for the different methods. Free Mass, Fat Mass and Body Weight in cirrhotic patients. The main advantage of DXA is that regional measure- ments may be made (for each arm, leg, the head and trunk), Results which are particularly useful in bone mass estimation. In this study, DXA was applied to a group of cirrhotic Regional and whole body composition analysis in the entire patients without overt ¯uid retention. This does not exclude group of cirrhotics and controls is reported in Table 3. A slight over-hydration in our patients. However it has been signi®cant reduction in whole body bone mineral content, recently shown that slight changes in hydration have little bone mineral density and percent fat was observed in in¯uence on the estimation of soft tissue composition by cirrhotics. By dividing the subjects by sex (Tables 4 and DXA (Pietrobelli et al, 1996). The comparison with age

Table 4 Regional analysis of body composition in male cirrhotics and controls (Mean Æ s.d.)

BMC BMD FM FF-MFM Total Fat (kg) (g=cm2) (kg) (kg) (kg) ( % )

Cirrhotics 0.44 Æ 0.05 0.94 Æ 0.01 3.14 Æ 1.82 5.89 Æ 0.80 9.47 Æ 2.39 31.06 Æ 10.30 Arms Controls 0.53 Æ 0.05 1.04 Æ 0.01 3.87 Æ 2.03 7.36 Æ 0.72 11.77 Æ 2.31 31.54 Æ 9.30 t-test P < 0.005 P < 0.005 NS P < 0.001 NS NS

Cirrhotics 0.72 Æ 0.09 0.87 Æ 0.01 10.95 Æ 4.29 27.90 Æ 1.98 39.58 Æ 4.33 27.05 Æ 8.40 Trunk Controls 0.92 Æ 0.08 1.00 Æ 0.01 13.01 Æ 3.90 27.89 Æ 2.22 41.83 Æ 3.19 30.75 Æ 7.60 t-test P < 0.0005 P < 0.0001 NS NS NS NS

Cirrhotics 1.07 Æ 0.12 1.28 Æ 0.01 6.47 Æ 2.80 16.00 Æ 1.44 23.56 Æ 3.80 26.55 Æ 7.80 Legs Controls 1.34 Æ 0.13 1.41 Æ 0.01 6.31 Æ 1.58 19.34 Æ 1.87 27.00 Æ 2.39 23.31 Æ 4.90 t-test P < 0.0005 P < 0.005 NS P < 0.0005 P < 0.05 NS

Cirrhotics 2.73 Æ 0.26 1.15 Æ 0.01 21.81 Æ 8.53 53.03 Æ 3.18 77.58 Æ 9.86 27.27 Æ 7.70 Total body Controls 3.29 Æ 0.29 1.26 Æ 0.01 24.22 Æ 6.76 57.99 Æ 4.04 85.52 Æ 4.88 28.12 Æ 6.61 t-test P < 0.0005 P < 0.001 NS P < 0.01 NS NS

BMC ˆ Bone Mineral Content; BMD ˆ Bone Mineral Density; FM ˆ Fat Mass; FF MFM ˆ Fat-Free Mineral-Free Mass. Body composition analysis in cirrhotic patients O Riggio et al 813 Table 5 Regional analysis of body composition in female cirrhotics and controls (Mean Æ s.d.)

BMC BMD FM FF-MFM Total Fat (kg) (g=cm (kg) (kg) (kg) ( % )

Cirrhotics 0.26 Æ 0.07 0.76 Æ 0.10 3.57 Æ 2.79 3.93 Æ 1.44 7.77 Æ 4.25 41.80 Æ 11.2 Arms Controls 0.27 Æ 0.01 0.85 Æ 0.01 4.16 Æ 1.20 3.89 Æ 0.49 8.33 Æ 1.61 49.17 Æ 5.60 t-test NS P < 0.05 NS NS NS NS

Cirrhotics 0.48 Æ 0.11 0.75 Æ 0.80 9.20 Æ 4.88 20.22 Æ 3.39 29.91 Æ 7.29 29.41 Æ 8.70 Trunk Controls 0.68 Æ 0.17 0.89 Æ 0.01 13.55 Æ 3.01 19.09 Æ 2.74 33.33 Æ 5.47 40.37 Æ 3.90 t-test P < 0.01 P < 0.001 P < 0.05 NS NS P < 0.005

Cirrhotics 0.67 Æ 0.14 1.00 Æ 0.12 6.88 Æ 2.73 11.28 Æ 1.80 18.83 Æ 4.37 35.27 Æ 7.90 Legs Controls 0.80 Æ 0.06 1.14 Æ 0.01 7.62 Æ 2.46 11.39 Æ 1.05 19.83 Æ 3.13 37.69 Æ 6.70 t-test P < 0.05 P < 0.001 NS NS NS NS

Cirrhotics 1.84 Æ 0.32 1.00 Æ 0.01 20.36 Æ 10.10 38.19 Æ 6.14 60.40 Æ 15.36 32.16 Æ 8.40 Total body Controls 2.25 Æ 0.27 1.11 Æ 0.01 26.30 Æ 3.42 37.07 Æ 1.97 65.63 Æ 4.79 39.94 Æ 2.70 t-test P < 0.01 P < 0.01 NS NS NS P < 0.05

BMC ˆ Bone Mineral Content; BMD ˆ Bone Mineral Density; FM ˆ Fat Mass; FF-MFM ˆ Fat Free Mineral-Free Mass. and BMI-matched healthy controls shows that in cirrhotic the physical activity index, emphasising the role of lean patients soft tissue is characterised by a reduced percent of (muscle) tissue as an important factor for bone mainte- fat mass. This agreees with what has already been reported nance. Thus, factors in¯uencing muscle mass, such as using other methodologies, such as skinfold anthropometry nutritional depletion, altered protein turnover and physical (Madden & Morgan, 1994), TBK (Crawford et al, 1994) inactivity may also, at least in part, negatively affect bone and TBW (McCoullogh et al, 1991). However, if a distinc- density. Although these factors are probably not speci®c for tion is made between genders, reduced percent fat mass liver cirrhosis, they should be taken into consideration in was only evident in females. future studies on bone alteration in these patients. Gender differences in body tissue loss have already been reported in epidemiological studies on the prevalence of malnutrition in cirrhotic patients (Italian Multicentre Coop- References erative Project on nutrition in liver cirrhosis 1994; Caregaro et al, 1996). In such studies adipose tissue was more Aloia JF, Vaswani A, Ma R & Flaster E (1995): Comparative study of body composition by Dual Energy X-Ray Absorptiometry. J. Nucl. frequently depleted in females than in males with cirrhosis, Med. 36, 1392±1397. while lean tissue was more often affected in males. Also, in Bonkovsky H, Hawkins M, Steinberg K, Hersh T, Galambos J, Henderson the present study, fat-free mineral-free mass was signi®- J, Millikan W & Galloway J (1990): Prevalence and prediction of cantly lower, in absolute terms, in male cirrhotics only. osteopenia in chronic liver disease. Hepatology 12, 273±280. Thus, according to the present and previous observations, Bramley P, Oldroyd B, Stewart S, Simpson M, Truscott J, Losowsky M & Smith M (1993): Body composition analysis in liver cirrhosis. The two different patterns of tissue loss may be found in measurement of body fat by Dual Energy X-Ray Absorptiometry in cirrhotic patients: in women lean tissue is maintained comparison to skinfolds anthropometry, bioelectrical impedance and while fat stores are reduced, as in early starvation; in total body potassium. In composition. eds. KJ Kellis & JD men lean tissue is reduced and body fat is normal, as Eastman. New York: Plenum Press. 211±214 Caregaro L, Alberino F, Amodio P, Markel C, Bolognesi M, Angeli P & seen under conditions of stress. These two patterns of tissue Gatta A (1996): Malnutrition in alcoholic and virus-related cirrhosis. loss have also been observed in cirrhotics by Crawford et Am. J. Clin. Nutr. 63, 602±609. al, in 1994, who unfortunately did not provide gender Crawford DHG, Shepherd RW, Halliday JW, Cooksley GWGE, Golding differentiated data. The reason for a gender difference in SD, Cheng WSC & Powell LW (1994): Body composition in nonalco- tissue loss in cirrhotic patients could be merely related to holic cirrhosis: the effects of disease etiology and severity on nutritional compartments. Gastroenterology 106, 1611±1617. the abundance of fat stores in female, which are progres- De Lorenzo A, Barra PFA, Sasso GF, Battistini NC & Deurenberg P sively used to cope with the metabolic needs until muscle (1991): Body impedance measurement during dialysis. Eur. J. Clin. mass remains the exclusive energy store. Nutr. 45, 321±325. Frost GS & Corish C (1989): Reproducibility of upper-arm anthropometry in subjects of differing body mass. J. Human Nutr. Dietology 2, 403± Conclusions 406. Guglielmi FW, Contento F, Laddaga L, Panella C & Francavilla A (1991): Previous observations on reduced bone mineral content and Bioelectrical impedance analysis: experience with male patients with density in cirrhotic patients (Bonkovsky et al, 1990) are cirrhosis. Hepatology 13, 892±895. con®rmed in our study. Bone loss in patients with cirrhosis Italian Multicentre Cooperative Project on nutrition in liver cirrhosis (1994): Nutritional status in cirrhosis. J. Hepatol. 21, 317±325. of various etiology and not only in those affected by Jebb SA & Elia M (1991): Assessment of changes in total body water in alcoholic or cholestatic liver diseases has emerged as a patients undergoing renal dialysis using bioelectrical analysis. Proc. clinical problem, because bone alterations are not corrected Nutr. Soc. 50, 153A. but are even accentuated after liver transplantation (Meys Johansson AG, Forslund A, Sjodin A, Mallmin H, Hambraeus L & Ljunghall S (1993): Determination of body compositionÐa comparison et al, 1994). The mechanism underlying bone loss in of Dual Energy X-Ray Absorptiometry and hidrodensitometry. Am. J. cirrhotic patients is complex and multifactorial. One basic Clin. Nutr. 57, 323±326. assumption, however, is that skeletal mass and density are Katch FI, Soloman RT, Shayevitz M & Shayevitz B (1986): Validity of both in¯uenced by mechanical loading, namely gravita- bioelectrical impedance to estimate body composition in cardiac and tional forces and muscular activity (Slemenda, 1995). In the pulmonary patients. Am. J. Clin. Nutr. 43, 972±973. Lang P, Steiger P, Faulkner K, Gluer C & Genant HK (1991): Osteo- present study, signi®cant correlations were found in cirrho- porosis: current techniques and recent developments in quantitative tics between BMD and both fat-free mineral-free mass and bone densitometry. Radiol. Clin. North Am. 29, 49±76. Body composition analysis in cirrhotic patients O Riggio et al 814 Lukasky HC (1993): Soft tissue composition and bone mineral status: Pietrobelli A, Formica C, Wang Z & Heyms®eld SB (1996): Dual-energy evaluation by Dual-Energy X-Ray Absorptiometry. J. Nutr. 123, 438± X-ray absorptiometry body composition model:review of physical 443. concepts. Am. J. Physiol. 271, E941±E951. Madden AM & Morgan MY (1994): A comparison of skinfold Podolsky S, Zimmerman HJ, Burrows BA, Cardarelli JA & Pattavina CG anthropometry and bioelectrical impedance analysis for measuring (1973): Potassium depletion in hepatic cirrhosis. A reversible cause of percentage body fat in patients with cirrhosis. J. Hepatol. 21, 878± impaired growth-hormone and insulin response to stimulation. New 883. England J. Med. 288, 644±648. Mazess RB, Bardin HS, Bisek JP & Hanson J (1990): Dual energy x-ray Prijatmoko DWI, Strauss BJG, Lambert JR, Sievert W, Stroud DB, absorptiometry for total body and regional bone mineral and soft tissue Wahlqvist ML, Katz B, Colman J, Jones P & Korman MG (1993): composition. Am. J. Clin. Nutr. 51, 1106±1112. Early detection of protein depletion in alcoholic cirrhosis: role of body McCullough AJ, Mullen KD & Kalhan SC (1991): Measurements of total composition analysis. Gastroenterology 105, 1839±1845. body and extracellular water in cirrhotic patients with and without Pritchard JE, Nowson CA, Strauss BJ, Carlson JS, Kaymakci B & Wark ascites. Hepatology 14, 1102±1111. JD (1993): Evaluation of Dual Energy X-Ray Absorptiometry as a Merli M, Riggio O, Dally L & PINC (1996): Does malnutrition affect method of measurement of body fat. Eur. J. Clin. Nutr. 47, 216±228. survival in cirrhosis. Hepatology 23, 1041±1046 Slemenda CW (1995): Editorial: Body composition and skeletal densityÐ Meys E, Fontages E, Fourcade N, Thomasson A, Pouyet M & Delmas PD Mechanical loading or something more? J. Clin. End. Met. 80, 1761± (1994): Bone loss after orthotopic liver transplantation. Am. J. Med. 97, 1763. 445±450. Svendsen OL, Haarbo J, Hassager C & Christiansen C (1993): Accuracy of Muller MJ, Lautz HU, Plogmann B, Burger M, Korber J & Schmidt FW measurements of body composition by Dual Energy X-Ray Absorptio- (1992): Energy expenditure and substrate oxidation in patients with metry in vivo. Am. J. Clin. Nutr. 57, 605±608. cirrhosis: the impact of cause, clinical staging and nutritional state. Wilson PWF, Paffenbarger RS, Morris JN & Havlink RJ (1986): Assess- Hepatology 15, 782±794. ment methods for physical activity and physical ®tness in population Olroyd B, Bramley PN, Stewart SP, Simpson M, Truscott JC, Losowsky studies: Report of a NHLBI workshop. Am. Heart J. 111, 1177±1192. MS & Smith MA (1993): A four-compartment model to determine body composition in liver cirrhosis. In Human Body Composition, eds. KJ Kellis & JD Eastman. New York: Plenum Press. 221±224