estimates of body water, fat-free mass, and body fat in patients on peritoneal dialysis by...

Kidney International, Vol. 63 (2003), pp. 1605–1617

PERSPECTIVES IN RENAL MEDICINE

Estimates of body water, fat-free mass, and body fat inpatients on peritoneal dialysis by anthropometric formulas

ANTONIOS H. TZAMALOUKAS, GLEN H. MURATA, DOROTHY J. VANDERJAGT, and ROBERT H. GLEW

Medicine Service, New Mexico Veterans Affairs Health Care System and University of New Mexico School of Medicine,Albuquerque, New Mexico; and Department of Biochemistry and Molecular Biology, University of New Mexico School ofMedicine, Albuquerque, New Mexico

Estimates of body water, fat-free mass, and body fat in patients estimates of Kt/V urea [1, 2]. In addition to their impor-on peritoneal dialysis by anthropometric formulas. tance in evaluating the adequacy of peritoneal dialysis

Background. Anthropometric formulas that are used to esti- by urea kinetics, estimates of V are useful in elucidatingmate body water in peritoneal dialysis patients can also bethe mechanism of hypertension [3], monitoring dry weightused to estimate fat-free mass and body fat. Evaluation of bodyand estimating body cell mass [abstract; Dumler F, Peritcomposition by the anthropometric formulas rests on two as-

sumptions: (1) fat contains no water, and (2) the water content Dial Int 20 (Suppl 1):S21, 2000], and evaluating intracel-of the fat-free mass is constant (72%). lular versus extracellular distribution of body fluids [ab-

Methods. We compared estimates of body water, fat-freestracts; Oe B et al, Perit Dial Int 20 (Suppl 1:S64, 2000;mass, and body fat by anthropometric formulas to estimatesPlum J et al, Perit Dial Int 20 (Suppl 1):S26, 2000].employing dilution of tracer substances to measure body water

and standard methods to analyze body composition in studies Clinical practice guidelines recommend the use of an-performed on peritoneal dialysis patients. We also analyzed the thropometric formulas for estimating V [4, 5]. These for-potential errors of the estimates of body composition by the mulas calculate V as a function of body weight (W) andformulas.

height (H). In addition, age (A), diabetes (D), and eth-Results. Estimates of the average body composition providednicity are also determinants of V in some formulas. Allby the anthropometric formulas agreed with estimates provided

by the standard methods. However, these formulas have the formulas take into account the degree of obesity, whichpotential of introducing large errors when estimating body is a major cause of discrepancy between estimates of V bycomposition in individuals differing from the average subject, different methods [6, 7]. This is important in peritonealeither because the anthropometric formulas do not account

dialysis, because peritoneal dialysis populations usuallyfor major determinants of body composition, such as physicalcontain large numbers of obese subjects [8]. In subjectsexercise, nutrition, and catabolic illness, or because these for-

mulas systematically overestimate body water in subjects who developing obesity, V increases but the body water con-are obese or experiencing volume excess. tent (the fraction V/W) decreases [9]. All formulas calcu-

Conclusion. Anthropometric formulas currently in existence late increasing V and decreasing V/W in subjects whocan provide only approximations of body composition and mayare gaining weight (see below).be the sources of large errors in evaluating body composition

The guidelines proposed formulas estimating V in peri-in peritoneal dialysis patients. The potential errors includeestimates of body water. These errors may alter the interpreta- toneal dialysis that were derived from subjects withouttion of urea kinetic studies in certain categories of peritoneal renal failure by comparing demographic data (gender,dialysis patients (e.g., obese subjects). age) and anthropometric measurements (height, weight)

to estimates of body water obtained by standard methods(usually isotopic water dilution techniques). Recently,Estimates of body water (V) have multiple applica-Johansson et al [10] produced the first anthropometrictions in patients on peritoneal dialysis. Their applicationformulas developed in a peritoneal dialysis population.in urea kinetic analysis is well known. Different methodsTable 1 shows the Sahlgrenska formulas, as Johansson et alestimating V may cause substantial differences in thenamed their formulas, and the Watson, Watson, and Butt[11], Hume and Weyers [12], and Chumlea et al [13]

Key words: anthropometric formula, peritoneal dialysis. formulas, which were developed in subjects without renalfailure, as well as the formula of Lee et al [14], which

Received for publication April 19, 2002was developed in hemodialysis patients by comparingand in revised form July 11, 2002

Accepted for publication November 18, 2002 anthropometric measurements to bioelectric impedanceanalysis (BIA) postdialysis. Table 2 shows the formula of 2003 by the International Society of Nephrology

1605

Tzamaloukas et al: Body water in peritoneal dialysis1606

Table 1. Anthropometric formulas for estimating body water (V)

Weight Height Age Body mass indexReference Y-intercept kg cm years kg/m2

MenSahlgrenska [10] �10.759 0.312 0.192 �0.078 —Watson, Watson, and Butt [11] 2.447 0.3362 0.1074 �0.09516 —Hume and Weyers [12] �14.012934 0.296785 0.194786 — —Chumlea et al, Caucasian [13] 23.04 0.50a — �0.03 �0.62Chumlea et al, African American [13] �18.37 0.34 0.25 �0.09 —Lee et al [14] �28.3497 0.366248 0.243057 — —

WomenSahlgrenska [10] �29.994 0.214 0.294 �0.0004 —Watson, Watson, and Butt [11] �2.097 0.2466 0.1069 — —Hume and Weyers [12] �35.270121 0.183809 0.344547 — —Chumlea et al, Caucasian [13] �10.50 0.20 0.18 �0.01 —Chumlea et al, African American [13] �16.71 0.22 0.24 �0.05 —Lee et al [14] �26.6224 0.232948 0.262513 — —

Both gendersSahlgrenska [10] �42.789 0.274 0.372 �0.033 —a The formula of Chumlea et al for caucasian men defines a linear relationship between body water (V) and body weight (W). The slope of this relationship

depends on height (H) and can be determined from the formula V � 23.04 � 0.50 W – 0.62 BMI � 0.03 A, where A is age. This formula can be rewritten as V �23.04 � 0.03A � [(0.50H2 � 0.62)/H2] W, where the expression (0.50H2 � 0.62)/H2 represents the V/W slope (coefficient �). This coefficient increases with height.For example, the coefficient � is 0.184 L/kg for a height of 140 cm, 0.285 L/kg for a height of 170 cm, and 0.345 L/kg for a height of 200 cm.

Table 2. The formula of Chertow et al for estimating where V is in liters, weight (W) in kilograms and heightbody water prehemodialysis

(H) in centimeters.Variable Coefficient Because the V/W ratio reflects the degree of obesity [9],Weight kg �0.04012056 body water estimates have been used to evaluate bodyHeight cm 0.12703384 composition. The formulas that estimate V also provideAge years �0.07493713

estimates of fat-free mass (FFM) and body fat (BF). TheGender (F � 0, M � 1) �1.01767992Diabetes (No � 0, yes � 1) 0.57894981 purpose of this report is to present the current status ofWeight2 �0.00067247 the use of anthropometric formulas for estimating V inAge � gender �0.03486146

the analysis of body composition in peritoneal dialysis.Gender � weight 0.11262857Age � weight 0.00104135 We provide an analysis of the relationships between V,Height � weight 0.00186104 FFM, and BF that are obtained using these formulas.

We also discuss the limitations and potential errors in-herent in these formulas when they are used to estimate

Chertow et al [15], which was developed in hemodialysis body composition in peritoneal dialysis.patients by comparing demographic data and anthropo-metric measurements to BIA predialysis. In addition to

BODY COMPOSITION ANALYSIS BASED ONcoefficients for weight, height, and age, that are includedANTHROPOMETRIC FORMULASin the formulas of Table 1, the formula of Chertow et

al includes coefficients for diabetes and for interactions All of the formulas shown in Table 1 assume a linearbetween age, gender, height, and weight. relationship between a change in body water (�V) and

The formulas in Tables 1 and 2 estimate V in adult sub- the corresponding change in body weight (�W). The co-jects. The Mellits and Cheek [16] formulas estimating V efficient �, assigned to weight [for example, 0.312 L/kgin children are listed below: in the Sahlgrenska formula for men (Johansson et al)],

represents this constant ratio �V/�W. Note that the coef-For boys with height �132.7 cm:ficient � is higher in men than women, regardless of the

V � �1.927 � 0.465 W � 0.045 H [Eq. 1] formula used (in the case of the Chumlea et al formulafor Caucasian men, the coefficient � becomes lower thanFor boys with height �132.7 cm:the highest coefficient � for women, that of the Watson,

V � �21.993 � 0.406 W � 0.209 H [Eq. 2]Watson, and Butt formula, at heights 156 cm (see foot-

For girls with height �110.8 cm: note of Table 1).For a given subject in whom age and height are con-V � 0.076 � 0.507 W � 0.013 H [Eq. 3]

stant, each formula calculates a unique (maximal) weightFor girls with height �110.8 cm: at which the body fat content is zero (WBF�0). The con-

cept of the value WBF�0 is not necessary for estimatingV � �10.313 � 0.252 W � 0.154 H [Eq. 4]

Tzamaloukas et al: Body water in peritoneal dialysis 1607

Table 3. Body weight at zero body fat (WBF � 0) and corresponding body mass index (BMI) values predicted by the anthropometric formulasestimating body water for 50-year-old subjects with normal hydration (body water is 72% of body weight)

Height � 170 cm Height � 140 cm

Reference WBF � 0, kg BMI, kg/m2 WBF � 0, kg BMI, kg/m2

MenSahlgrenskaa [10] 44.1 15.2 29.3 15.3Watson, Watson, and Butta [11] 41.6 14.4 33.2 16.9Hume and Weyers [12] 45.1 15.6 31.3 16.0Chumlea et al, Caucasiana [13] 49.6 17.2 40.2 20.5Chumlea et al, African Americana [13] 51.7 17.9 31.9 16.3Lee et al [14] 36.7 12.7 16.1 8.2Chertow et al, NDa [15] 48.4 16.8 31.7 16.2Chertow et al, Da [15] 50.1 17.3 33.2 16.9

WomenSahlgrenskaa [10] 39.5 13.7 22.0 11.2Watson, Watson, and Butt [11] 34.0 11.8 27.2 13.9Hume and Weyers [12] 43.5 15.0 24.2 12.3Chumlea et al, Caucasiana [13] 37.7 13.0 27.3 13.9Chumlea et al, African Americana [13] 43.2 14.9 28.9 14.7Lee et al [14] 37.0 12.8 20.8 10.6Chertow et al, NDa [15] 42.5 14.6 30.0 15.3Chertow et al, Da [15] 43.8 15.1 31.2 15.9

Both gendersSahlgrenskaa [10] 42.0 14.5 16.9 8.6

Abbreviations are: ND, nondiabetic subjects; D, subjects with diabetes mellitus.a Values of WBF � 0 and their corresponding BMI decrease as age increases

body composition by the anthropometric formulas, but 11.9 kg (BMI 9.8 kg/m2); for boys 140 cm tall, 23.1 kg(BMI 11.8 kg/m2); for girls 110 cm tall, 7.1 kg (BMI 5.8is useful when considering the limitations of these formu-kg/m2); and for girls 140 cm tall, 24.0 kg (BMI 12.3 kg/m2).las. In addition, the concept of WBF�0 offers a theoretical

The average subjects in the studies represented in thedefinition of wasting (weights below the value WBF�0).formulas of Table 1 had no hydration abnormalities.The value WBF�0 can be calculated from a formula byBy these formulas, therefore, FFM � WBF�0 when bodyconsidering that body weight consists of two parts, FFMweight is at WBF�0. In contrast, the average subjects inand BF (W � FFM � BF). Body water is contained inthe formula of Chertow et al (Table 2) had volume excessits entirety in the FFM. In subjects without hydration[15, 19]. Therefore, when body weight is at WBF�0 by theabnormalities, the water content of the FFM is withinChertow et al formula, FFM WBF�0 [20].very narrow limits and at first approximation can be con-

For weights below the WBF�0, values calculated by thesidered as constant [17, 18]. We assume in this analysisanthropometric formulas, as indicated in equation 5, thethat V/FFM equals 0.72. In a subject without volume ab-formulas produce V/W values greater than 0.72, andnormalities and with V/W � 0.72, FFM � W and BF � 0suggest consequently that all subjects with wasting are(the subject is at WBF�0). Therefore, the value WBF�0 canoverhydrated. For example, body water is 28.9 L by thebe calculated for each formula and each set of constantSahlgrenska formula (by Johansson et al) in a 50-year-conditions (height, age), by calculating the weight atold man with a height of 170 cm and a weight of 35 kg,which V/W � 0.72. For example, if V � 0.72 WBF�0, thewhich is below the WBF�0 value of 44.1 kg reported inSahlgrenska formula (by Johansson et al) for a 50-year-Table 3. The V/W ratio is 28.9/35 or 0.83, suggesting over-old man with a height of 170 cm can be restated as:hydration. This may represent a systematic error if wast-

0.72 WBF�0 � 0.312 WBF�0 � 0.192 � 170 ing does not systematically affect the hydration of FFM.This is the reason why we considered the value WBF�0 as� 0.078 � 50 � 10.759 (Eq. 5)being unique. We suggest that the anthropometric formu-

or, WBF�0 � 44.1 kg. For a 20-year-old man with the las are not applicable at weights below the value WBF�0.same height, the Sahlgrenska formula (by Johansson et If a person at WBF�0 starts gaining weight from obesity,al) for men estimates a WBF�0 equal to 49.7 kg. Table 3 some of the increase in weight (�W) will be due to fatshows WBF�0 values and corresponding body mass index (�BF) and the other part will be reflective of an in-(BMI) values for 50-year old subjects with height equal crease in FFM (�FFM). The fat component of the weightto 170 or 140 cm calculated from the formulas in Table 1 gain will not contain water. All of the increase in bodyand Table 2. The Mellits and Cheek formulas calculate the water (�V) will be in the newly acquired FFM, which

will have water content equal to 0.72 (�V � 0.72�FFM).following WBF�0 values for children: for boys 110 cm tall,


Table 4. Changes in fat-free mass (�FFM) and in body fat (�BF)as a function of weight increase (�W) above the WBF � 0 values

predicted by anthropometric formulas assuming linear relationships

�FFM/�W �BF/�WReference kg/kg kg/kg

MenSahlgrenska [10] 0.433 0.567Watson, Watson, and Butt [11] 0.467 0.533Hume and Weyers [12] 0.412 0.588Chumlea et al, Caucasiana [13] 0.396 0.604Chumlea et al, African American [13] 0.472 0.528Lee et al [14] 0.509 0.491

WomenSahlgrenska [10] 0.297 0.703Watson, Watson, and Butt [11] 0.343 0.657Hume and Weyers [12] 0.255 0.745Chumlea et al, Caucasian [13] 0.278 0.722Chumlea et al, African American [13] 0.306 0.694Lee et al [14] 0.324 0.676

Both gendersSahlgrenska [10] 0.381 0.619a The values �FFM/�W and �BF for Caucasian men are height dependent.

The value calculated for this Table was for a height of 170 cm. �FFM/�W valuesfor a height of 140 cm are 0.256 and 0.744 kg/kg, respectively. Correspondingvalues for a height of 200 cm are 0.479 and 0.521 kg/kg, respectively (see alsofootnote of Table 1).

From the anthropometric formula, �V � ��W, wherethe coefficient � for weight can be obtained from Ta-ble 1. Therefore, Fig. 1. Fat-free mass (FFM) in kg, body water (V) in L, and body fat

(BF) in kg, in a man (A) and a woman (B), both 50 years old and 1700.72 �FFM � ��W (Eq. 6) cm tall, at body weights between the WBF�0 value and the value at bodymass index (BMI) � 50 kg/m2. FFM, V, and BF were calculated by the

or use of the Sahlgrenska formulas (as reported by Johansson et al [10]).

�FFM � ��W/0.72 (Eq. 7)

and Table 4 shows the �FFM/�W and �BF/�W valuesfor weights exceeding the WBF�0 value predicted by the

�BF � �W (1 � �/0.72) (Eq. 8)formulas of Table 1. All formulas predict a greater rate

For subjects with W � WBF�0, FFM will be the sum of of increase in BF in women than men having the sameWBF�0 and the value shown in equation 3, or increase in body weight above the WBF�0 value. In popu-

lations without renal failure developing obesity, the ratioFFM � WBF�0 � ��W/0.72 (Eq. 9)

�FFM/�W is, indeed, greater in men than in women[21]. Figure 1 shows body water (V), FFM, and BF in awhile body fat will be obtained by equation 8. Usually,50-year old man and a 50-year-old woman, both with aFFM is estimated as FFM � V/0.72, where V is the bodyheight of 170 cm, whose BMI fluctuates between thewater value found by the formula. This estimate of FFMBMI at WBF�0 by the Sahlgrenska formulas (by Johans-and that obtained from equation 9 are equal. Also, BFson et al) (Table 3) and 50 kg/m2.is usually estimated as BF � W – FFM, and this estimate

Figure 2 shows the fractions V/W, FFM/W, and BF/Wis equal to the BF estimate obtained using equation 8.in the subjects presented in Figure 1. With progressiveConsider the following example using the Sahlgrenskaweight gain from obesity, the fractions V/W and FFM/W(by Johansson et al) formula for men. A 50-year-old mandecrease, while the fraction BF/W increases progres-with a height of 170 cm and a WBF�0 value of 44.1 kgsively. At any given weight W � WBF�0, the fraction V/W(Table 3), with actual weight equal to 100 kg, has V �will be between the values 0.72 and � (in Fig. 2, the co-49.2 L. Therefore, FFM is 49.2/0.72 � 68.3 kg and BF isefficient � is 0.312 L/kg in the man and 0.214 L/kg in100 � 68.3 � 31.7 kg. The difference between actualthe woman). At the weight W, body water is the sum ofweight (W) and WBF�0 (�W) is 100 � 44.1 � 55.9 kg.body water at WBF�0 and the water of �W, orAccording to equation 7, �FFM � 0.312 � 55.9/0.72 �

24.2 kg. According to equation 9, FFM � 44.1 � 24.2 �V � 0.72 WBF�0 � ��W (Eq. 10)

68.3 kg. According to equation 8, BF � 55.9 (1 – 0.312/0.72) � 31.7 kg. In the example of the 50-year-old man who weighs


could not calculate average body composition. For allcalculations in Table 5, we assumed a V/FFM ratio of0.72. This caused small variations from the reported bodycomposition in studies, which found average V/FFM val-ues that differed slightly from 0.72. In our analysis, weused the Chumlea et al [13] formulas for Caucasian sub-jects for all studies, except the study of Dahl et al [25],for which we used the Chumlea formulas for both Cauca-sian and African American subjects in view of the largenumbers of African Americans with renal failure in theirstudy population in New Jersey.

The standard methods used to measure body waterin the studies presented in Table 5 included tritiatedwater dilution [10], stable H2

18O measurements [22], anti-pyrine dilution [23], D2O dilution [24], and ethanol dilu-tion [25]. In addition to showing the standard methodsthat were used for estimating body water and the esti-mates we arrived at from the anthropometric formulas,Table 5 also shows estimates of body composition arrivedat by the authors of the respective studies using alterna-tive methods to analyze body composition. These meth-ods include skin fold thickness [22, 23], which is an es-tablished method for analysis of body composition [29]and one that has been used to evaluate nutrition in longi-tudinal studies of peritoneal dialysis patients [30], BIA[22–24], which has been used extensively to analyze body

Fig. 2. The fractions fat-free mass over body weight (FFM/W), body composition in hemodialysis [31–44], and peritoneal di-water over body weight (V/W), and body fat over body weight (BF/W) alysis [20, 45–60] patients, dual-energy x-ray absorpti-in a man (A) and a woman (B), both 50 years old and 170 cm tall, at

ometry (DEXA) [24], which is considered as the goldbody weights between the WBF�0 value and the value at body massindex (BMI) � 50 kg/m2. FFM, V, and BF were calculated by the standard for analysis of certain elements of body com-Sahlgrenska formulas (as reported by Johansson et al [10]). position, such as bone mass and body fat, in peritoneal

dialysis [10, 61–64], and creatinine kinetics [23], whichwas proposed as a means of estimating FFM [65] and100 kg, is 170 cm tall, and has a WBF�0 of 44.1 kg accordinghas been used to analyze body composition in peritonealto the Sahlgrenska formula (by Johansson et al), equa-dialysis [48].tion 10 yields the value V � 0.72 � 44.1 � 0.312 (100 �

Table 5 shows that the various anthropometric formu-44.1) � 49.2 L. Body water content (the value V/W) inlas do provide reasonable agreement with body composi-this case is 0.492 L/kg. Application of the Sahlgrenskation estimates obtained using the standard methods offormula (by Johansson et al) for men to the same setmeasuring body water. Estimates by DEXA and skinof data provides exactly the same V and V/W values,fold thickness (which is also an anthropometric methodrespectively.for estimating body composition that should not be con-In subjects without volume disorders, the relationshipfused with the anthropometric formulas) were also rea-between V, FFM, and BF is such that measurement ofsonably close. BIA estimates of V/W and FFM/W wereany one of the three provides estimates for the othersubstantially lower than the estimates arrived at by stabletwo. Several studies that analyzed body composition inH2

18O dilution in one study [22] and larger than theperitoneal dialysis applied this principle, which is basedestimates obtained by the use of D2O dilution in anotheron a fixed body water content in FFM (V/FFM � 0.72)study [24]. In a third study [23], BIA estimates of bodyand the absence of water in fat [22–25]. Table 5 showscomposition varied widely depending on the predictivethe results of the analysis of body composition conductedequations used.using the formulas shown in Table 1 in patients undergo-

Table 6 shows determinants of body composition (age,ing peritoneal dialysis, from studies which measuredweight, and height) and degree of obesity (BMI) of thebody water by a standard method [10, 22–25]. In thissubjects presented in Table 5 [10, 22–25]. The range oftable, we calculated body composition for the averagethe mean BMI values in these studies was narrow (24.0patient in each study. For studies in which height, weight,

or gender of the subjects were not reported [26–28], we to 27.1 kg/m2).


Table 5. Body water/body weight (V/W), fat-free mass/body weight (FFM/W), and body fat/body weight (BF/W)for the average subjects of studies of body composition in peritoneal dialysis

Watson,Watson, Hume and Chumlea

Reference Standarda Sahlgrenska and Butt Weyers et alb Lee et al Other

Johansson et al [10]Men V/W 0.554c 0.561 0.549 0.569 0.586 0.559 0.562d

FFM/W 0.770 0.779 0.762 0.790 0.803 0.776 0.781BF/W 0.230 0.221 0.238 0.210 0.197 0.224 0.219

Women V/W 0.492c 0.494 0.487 0.511 0.488 0.482 0.509d

FFM/W 0.683 0.688 0.677 0.710 0.678 0.669 0.708BF/W 0.317 0.313 0.323 0.290 0.322 0.331 0.291

Arkouchee [22] V/W 0.491f 0.495 0.503 0.513 0.515 0.488 0.361g

FFM/W 0.682 0.688 0.699 0.712 0.715 0.679 0.501BF/W 0.318 0.313 0.301 0.288 0.285 0.321 0.499

De Fijterh [23] V/W 0.541i 0.530 0.522 0.545 0.539 0.528 0.521g

FFM/W 0.752 0.736 0.725 0.757 0.749 0.733 0.724BF/W 0.248 0.264 0.275 0.243 0.251 0.267 0.276

Woodrowj [24] V/W 0.495k 0.522 0.520 0.535 0.535 0.509 0.543g

FFM/W 0.687 0.725 0.723 0.742 0.742 0.708 0.755BF/W 0.313 0.276 0.277 0.258 0.258 0.292 0.245

Dahll [25] V/W 0.530m 0.494 0.489 0.510 0.523 0.497 —FFM/W 0.736 0.686 0.679 0.709 0.727 0.690 —BF/W 0.264 0.314 0.321 0.291 0.273 0.310 —

a Standard reference method measuring body waterb Chumlea et al formula for Caucasiansc THOd Total body potassium countinge D2O, V/W � 0.512, FFM/W � 0.711, BF/W � 0.289; skin fold thickness, V/W � 0.518, FFM/W � 0.720, BF/W � 0.280f 18Og Bioelectrical impedance analysis (BIA)h Skin fold thickness, V/W � 0.514, FFM/W � 0.713, BF/W � 0.287; creatinine kinetics, V/W � 0.475, FFM/W � 0.660, BF/W � 0.340; BIA estimates by different

equations estimating FFM provided the following ranges: V/W 0.476–0.554, FFM/W 0.662–0.769, BF/W 0.231–0.328i Antipyrinej DEXA, V/W � 0.507, FFM/W � 0.704, BF/W � 0.296k D2Ol Chumlea et al formulas for African Americans, V/W � 0.539, FFM/W � 0.736, BF/W � 0.264.m Ethanol dilution

Table 6. Determinants of body composition in patients on peritoneal dialysis reported in Table 5

Body mass index (BMI)Reference Age years Weight kg Height cm kg/m2

Johansson et al, men [10] 57.814.5 74.512.0 176a 24.0 3.4Johansson et al, women [10] 56.714.4 62.814.1 161a 24.1 5.3Arkouche et al [22] 699 6613 158 9 26.5 5.0de Fijter et al [23] 626 75.6 10.9 1748 24.9 3.5Woodrow et al [24] 57.416.3 65.915.3 163.39.5 24.7b

Dahl et al [25] 52.317.5 76.016.7 1677 27.1 5.5

Values are presented as mean standard deviation.a Mean value calculated from mean weight and mean BMIb Mean value calculated from mean weight and mean height

LIMITATIONS AND POTENTIAL ERRORS OF conditions, which can result in systematic errors in esti-THE ANTHROPOMETRIC FORMULAS mates of body composition parameters.

The major limitation of the anthropometric formulasThe reasonable agreement between the estimates ofis that for each gender, age, height, and weight (and forbody composition provided by the anthropometric for-some formulas, even ethnicity) they compute a uniquemulas and standard methods for analyzing body compo-body composition, which is that of the average subject ofsition shown in Table 5 was achieved only for patientsthe populations in whom these formulae were developed.with an average body composition. Estimates of bodyThis problem is underscored by the fact that each for-composition by anthropometric formulas for subjectsmula computes a unique value of WBF�0 for each heightwho differ from the average subject may vary greatly(Table 3). This unique value reflects the body composi-from the actual body composition. In this section, we dis-

cuss the sources of the limitations of the formulas and tion of a subject with average body composition. In real-


ity, the weight at which there is no body fat should berepresented not by a unique value, but by a range ofvalues with the WBF�0 calculated by the anthropometricformula as the average value.

There are two reasons why WBF�0 should not be repre-sented by a single value for each gender, height andethnicity. First, the water content of FFM varies fromperson to person and is not a constant value (0.72), evenin subjects without a hydration disorder. In healthy sub-jects, age affects the V/FFM ratio, with prepubescentchildren and older adults having higher V/FFM valuesthan young adults [66]. A V/FFM range of 0.69 to 0.77was proposed for subjects without hydration disorders[67]. In a study that used D2O to measure body waterand DEXA to measure fat-free mass, Woodrow et al[63] found that V/FFM ratios ranged from 0.65 to 0.87in peritoneal dialysis patients. The high values in thisstudy were undoubtedly seen in subjects with volumeexcess. Nevertheless, using the Sahlgrenska formula (byJohansson et al) for males, the range of WBF�0 for a50-year-old man 170 cm tall (Table 3) would be 32.2 to53.2 kg (BMI 11.2 to 18.4 kg/m2) if the water content ofthe FFM varied between 0.65 and 0.87.

Figure 3 shows FFM and BF in two 50-year-old men,170 cm tall, with weights between the WBF�0 value andthat at a BMI of 50 kg/m2, with a �V/�W of 0.312 L/kgand a ratio V/FFM either 0.65 or 0.87 L/kg. The value of Fig. 3. Fat-free mass (FFM) (A) and body fat (BF) (B) in two men, both 50

years old and 170 cm tall, who have the same body water by the SahlgrenskaV is the same at the same weight, regardless of the ratioformula for men (as reported by Johansson et al [10]), at body weightsV/FFM. However, the ratio V/FFM affects greatly the between the WBF�0 value and that at body mass index (BMI) � 50 kg/m2

calculations of FFM and BF because of the different and with hydration of fat-free mass (the fraction V/FFM) either 0.65 or 0.87.WBF�0 values. For example, in a 100 kg man with thesame age and height as those of the man depicted inFigure 3, V will be 49.2 L regardless of the ratio V/FFM.

in an Asian population. There are no anthropometricHowever, FFM will be 75.7 kg if V/FFM is 0.65 and 56.5 kgformulas for estimating body water in other populations,if V/FFM is 0.87. Corresponding values of BF are 24.3including other Asian populations, African populations,and 43.5 kg, respectively. Therefore, in settings where theSouthern European populations, South American popu-V/FFM ratio varies, the anthropometric formula requireslations, and Native American populations. Other deter-an independent measure of the value V/FFM to arriveminants of body composition, including nutrition, levelat appropriate estimates of BF and FFM.of physical activity, skeletal frame [68], and catabolicThe current anthropometric formulas have a secondillness, particularly conditions that lead to muscle wast-major limitation. Important determinants of body com-ing, are absent from these formulas.position are incompletely analyzed or not included at all

All of the factors mentioned above can affect the FFMin these formulas. Determinants of body composition in-and consequently the value WBF�0 and are the main rea-corporated in these formulas include gender, age, height,sons for the variation in body water from the value pre-and weight. Because body composition differs substan-dicted by the regression reported in the original studiestially between the genders, formulas not accounting for[10–15]. The measure of this variation differed amongthe gender of the subjects, such as the Sahlgrenska for-the original studies reporting these formulas. The Sahl-mula (by Johansson et al) for both genders [10], are atgrenska report by Johansson et al [10] did not containleast theoretically prone to errors. Ethnicity is a majorany measure of variation. Watson, Watson, and Butt [11]determinant of body composition [13], but one that hasreported a standard deviation of the body water estimatebeen incompletely considered. The Sahlgrenska formu-equal to 3.76 L in men and 3.60 L in women. The studylas by Johansson et al [10] were developed in Scandina-by Hume and Weyers [12] reported complex formulasvian populations, while Chumlea et al [13] developed for-for estimating the standard error of the body water esti-mulas specific for African American and Caucasian Amer-

ican populations. The formula by Lee et al [14] was derived mates in men and women. The standard errors of the


Hume and Weyers formulas increase in both genders asheight or weight increases. For a man with a height of170 cm and a weight of 70 kg, we calculated a standarderror equal to 2.3 L. Chumlea et al [13] reported a rootmean square error between 3.8 and 5.0 L in men andbetween 3.3 and 3.6 L in women. Lee et al [14] did notreport a measure of variation in their study. Finally, Cher-tow et al [15] reported a root mean square error of 3.57 L.

Differences in body water from the values predictedby the anthropometric formulas would have significanteffects on the WBF�0 values presented in Table 3 evenwhen the ratio V/FFM is constant. Therefore, the WBF�0

values should be presented as a range of weights insteadof a single weight. For example, if body water variedwithin 1 standard deviation (3.76 L) of the mean valuecalculated by the formula of Watson, Watson, and Butt[11] for men, the range of WBF�0 for a 50-year-old manwith a height of 170 cm would be 31.8 to 51.3 kg andthe corresponding range of BMI values would be 11.0to 17.8 kg/m2 even if V/FFM is constant at 0.72.

An example of the effects of varying WBF�0 with con-stant hydration of FFM is provided by Figure 4, whichshows V, FFM, and BF for two 50-year-old men 170 cmtall with weight varying between the WBF�0 value andthat at a BMI of 50 kg/m2, in whom the value WBF�0 iseither 32.2 or 53.2 kg because of differences in otherdeterminants of FFM, but the ratio V/FFM is constantat 0.72 and the ratio �V/�W is the same as in the Sahl-grenska formula (by Johansson et al) for men (0.312L/kg). In this case, all three values characterizing bodycomposition (V, FFM, BF) differ. For example, if bodyweight is 100 kg and WBF�0 is 32.2 kg, V � 44.3 L, FFM �61.6 kg, and BF � 38.4 kg. If body weight is 100 kg, butWBF�0 is 53.2 kg, V � 52.9 L, FFM � 73.5 kg, and BF �26.5 kg. Therefore, in settings where the FFM deviatesfrom the norm because of the influence of one or more

Fig. 4. Body water (V) (A), fat-free mass (FFM) (B), and body fat (BF)of the conditions not accounted for in a formula, this (C ) in two men, both 50 years old and 170 cm tall, at body weights be-formula will yield values of V, FFM, and BF that may tween the WBF�0 value and that at body mass index (BMI) � 50 kg/m2.

The ratio �V/�W above the W BF�0 values were considered to be equalvary considerably from the actual values.to the value provided by the Sahlgrenska formula (as reported by Johans-The limitations of the anthropometric formulas refer son et al [10] for men (0.312 L/kg). We assumed that hydration of FFM

to conditions that may affect particular patients in a perito- was the same (V/FFM � 0.72) in both men, but that the two men hadtwo different WBF�0 values (32.2 and 53.2 kg, as indicated on the lines).neal dialysis population, but not the average patient.

Other conditions may cause systematic overestimates orunderestimates of V and FFM, with opposite results onBF, and could affect the average values calculated by trast, Wong et al [28] found lower V Watson, Watson, andthe anthropometric formulas. Several conditions may Butt [11] values than body water estimates by heavy watercause these problems, but only two, namely obesity and a dilution in obese peritoneal dialysis subjects, althoughhydration disorder, have been explored in the literature. the difference did not reach statistical significance. The

Johansson et al [10] reported higher V values by the patients in this last study did not have a normal hydrationWatson, Watson, and Butt [11] and Sahlgrenska (as re- status and this could have affected the results. It is likelyported by Johansson et al) formulas compared to those that the formulas in Table 1 overestimate body water inarrived at by tritiated water dilution analysis in obese the typical obese patient undergoing peritoneal dialysis.peritoneal dialysis subjects. In a study that involved rela- Investigation of the mechanisms of this overestimatetively few subjects, Woodrow et al [24] observed a similar could be helpful in identifying systematic defects of these

formulas. Three potential mechanisms are now discussed.trend, but it did not reach statistical significance. In con-


Obesity could affect FFM. As discussed earlier, FFMincreases as obesity-based body weight increases. It ispossible, however, that obesity causes a systematic de-crease in the value WBF�0. In the example of the 50-year-old man who is 170 cm tall, the Sahlgrenska formula (asreported by Johansson et al) predicts a WBF�0 value of44.1 kg (Table 3), and, in a 100 kg man of the same heightand age, a V of 49.2 L, a FFM of 68.3 kg, and a BF of31.7 kg. If obesity had decreased the value WBF�0 to 33.2kg, while the coefficient � (the ratio �V/�W) remainedthe same as in the Sahlgrenska formula (0.312 L/kg),actual body composition would be as follows (equations2 to 6): V � 44.7 L, FFM � 62.1 kg, and BF � 37.9 kg.The literature provides evidence that some peritonealdialysis patients lose FFM as they develop obesity [69, 70].

Fig. 5. Estimates of body water (�) obtained by the use of the formulaA second potential cause of high anthropometric Vby Chertow et al [17] in a 50-year-old nondiabetic man with a height

values in obese subjects is an inappropriately high � of 170 cm, at body weights between the WBF�0 value and that at bodymass index (BMI) � 50 kg/m2. The line shows linear regression ofcoefficient for weight in these formulas. It is interestingweight (W) on the body water estimates (V). The regression was asthat the Hume and Weyers formulas [12], which werefollows: V � 20.761 � 0.313W, r � 0.9984.

developed with special concern for obese subjects, con-tain the lowest values of the coefficient � for both menand women (in the case of the formula by Chumlea et alfor Caucasian men, this is true only for heights exceeding 140 and 145 kg. Despite this mathematical decrease, the

agreement between Chertow et al estimates of V and175 cm; see Table 1 and its footnote). If the sole mecha-nism of overestimation of body water by an anthropo- values predicted by the linear regression in Figure 5 is

extremely high (r 2 � 0.997).metric formula in obese subjects were an inappropriatelyhigh coefficient �, correction of the formulas would be A second finding of Figure 5 worth noticing is that

estimates of body water (V) from the linear regressionrelatively easy.A third mechanism that would account for the over- analysis are higher than the estimates of V from the

formula of Chertow et al in obese subjects. This findingestimate of body water by the formulas in Table 1 inobese peritoneal dialysis subjects is more complex than is in agreement with the findings of Johansson et al [10].

Estimates of V from the linear regression analysis inthe two discussed above. All of these formulas describelinear relationships between the increase in body weight lean individuals, even at weights exceeding slightly the

value WBF�0, also overestimate the Chertow et al V esti-and the corresponding increases in body water, FFM andBF (Fig. 1). It is quite possible that these relationships mates. This finding contrasts with the suggestion by Jo-

hansson et al [10] that the anthropometric formulas tendare, in fact, not linear (the coefficient � is not constant,but decreases with increasing weight). This possibility is to underestimate body water in lean individuals.

Finally, comparison of the regression shown in Fig-supported by the formula of Chertow et al (Table 2),which contains a negative coefficient for the square of ure 5 (V � 20.761 � 0.313 W) to the regression from

the Sahlgrenska formula (as reported by Johansson etthe weight. As a result, the ratio �V/�W calculated usingthis formula for weight gains is not constant, as in the al) for a man with the same age, height, and weight range

(V � 17.981 � 0.312 W) reveals that the two regressionsformulas shown in Table 1, but decreases progressivelyas weight increases. have the same coefficient �, indicating that they estimate

the same average change in body water if weight changesFigure 5 shows estimates by Chertow et al of bodywater in a 50-year-old man without diabetes mellitus between the boundaries set for these calculations in

50-year-old men with a height of 170 cm, while the y-inter-with a height of 170 cm at weights between the value atWBF�0 (48.4 kg, see Table 3) and that at a BMI of 50 cept of the linear regression derived from the Chertow

et al estimates of V is larger than that of the Sahlgrenskakg/m2, and the linear regression of weight on the Chertowet al V estimates. This figure illustrates the difficulty in formula. This larger y-intercept illustrates the fact that

the subjects in the study by Chertow et al had volumeappreciating errors caused by linear regression analysisof a mathematically curvilinear relationship under cer- excess while those in the study by Johansson et al did not.

Studies in populations without renal failure supporttain circumstances. The calculated formula of Chertowet al, �V/�W, decreases progressively with weight. For the notion of declining �FFM/�W ratio as obesity devel-

ops [71]. The progressive decrease in the �V/�W ratioexample, the average �V/�W is 0.370 L/kg for weightsbetween 50 and 55 kg and 0.249 L/kg for weights between and the �FFM/�W ratio with developing obesity will in-


evitably lead to systematic overestimation of body water an anthropometric formula, plus the difference betweenby anthropometric formulas that use linear �V/�W coef- actual and dry weight.ficients � in subjects who are either substantially more For example, a 50-year-old man without diabetes mel-obese or substantially leaner than the average patient. litus who is 170 cm tall and has a dry weight of 80 kg

The second potential source of systematic errors represents with an actual weight of 100 kg. The value ofsulting from the use of the anthropometric formulas in V in this man is 49.2 L according to the Sahlgrenskaperitoneal dialysis patients is an abnormality in hydra- formula (as reported by Johansson et al) (V/W � 0.492),tion. All of the formulas in Table 1, including the Sahl- and 52.5 L by the formula of Chertow et al (V/W �grenska formulas (developed by Johansson et al), were 0.525). At a dry weight of 80 kg, the value V by thedeveloped in subjects without hydration disorders. How- Sahlgrenska formula is 42.9 L (V/W � 0.537). The differ-ever, patients on peritoneal dialysis are often subject to ence between actual and dry weight (20 kg) consists ofmajor gains or losses of body water. Volume excess is a excess volume. Therefore, at a weight of 100 kg, V �common and well-known problem in peritoneal dialysis 42.9 � 20 � 62.9 kg (V/W � 0.629). Note that the body[72]. In one study, one fourth of the peritoneal dialysis water content (the value V/W) increases as the volumepopulation developed at least one episode of symptom- expands [77]. The Sahlgrenska formula, which deals ex-atic fluid retention during the course of peritoneal dial- clusively with weight gains due to developing obesity,ysis with an average fluid gain of 9.7 kg or 14% over the predicts a decreasing V/W as the weight increases fromdry weight [73]. Volume deficit, which also has impor- 80 to 100 kg. The same is true for all other formulas,tant clinical implications in peritoneal dialysis patients including the formula by Chertow et al, which, at a weight[74, 75], occurs less frequently than volume excess. In the of 80 kg, estimates a V of 46.0 L (V/W � 0.576) in thestudy by Aftentopoulos et al [75], only 2.6% of a large subject in this example.continuous ambulatory peritoneal dialysis (CAPD) pop- The previous discussion leads us to conclude that noulation developed hypotension secondary to volume formula that contains a reducing coefficient for weightdepletion. The anthropometric formulas have the poten- (a coefficient � 1.0) can avoid introducing systematictial of creating large systematic errors in the estimates of

errors in volume-expanded patients. These errors trans-body composition in subjects with volume disturbances.

late, in parallel, to errors in the estimates of FFM andThe V values calculated by the formulas are represen-BF. A formula estimating body water at dry weight,tative of subjects in whom there is no volume distur-with the correction presented for subjects with volumebance. The presence of volume excess in a peritonealdisturbances, would eliminate the systematic errors.dialysis population results in systematic underestimatesHowever, such a revised formula would itself depend onof V in the average subject when any of these formulasthe estimation of dry weight, which is prone to its ownare used [76]. The magnitude of this underestimate isset of errors.proportional to the degree of volume excess [77]. The

various formulas overestimate V in the average perito-neal dialysis patient with volume depletion [77]. CLINICAL IMPLICATIONS

The formula of Chertow et al (Table 2) provides valuesIn peritoneal dialysis subjects without volume distur-of V that are 9.5% higher, on the average, than the es-

bances, the formulas shown in Table 1 provide reliabletimates of V by Watson, Watson, and Butt [11] in PDestimates of body water, FFM, and BF for subjects withpatients [78]. The formula of Chertow et al will avoidaverage body composition. However, in individuals whosystematic deviations from the average value of V onlydeviate substantially from the average, estimates of bodyin subjects whose volume excess is 10% or less. A methodcomposition obtained using the anthropometric formulasestimating V appropriate for any degree of volume dis-may differ greatly from true values [10, 22–25]. This isturbance uses anthropometric formulas estimating V inparticularly true for obese peritoneal dialysis subjects forsubjects without volume disturbances, but requires knowl-whom these formulas systematically overestimate bodyedge of the subject’s dry weight [76].water and FFM and underestimate BF.According to this last method, body weight consists

In peritoneal dialysis subjects with volume disturbances,of dry weight and weight excess or deficit secondary tothe formulas shown in Tables 1 and 2 have the potentiala volume disturbance. Body water content at dry weightof introducing systematic errors into the estimates ofcan be calculated without systematic errors by any for-body composition of even an average subject, unless theymula that estimates V in the absence of volume distur-are corrected for the magnitude of the volume distur-bances (all of the formulas in Table 1). The differencebance. Such a correction requires an accurate assessmentbetween the actual and the dry weight is due entirely toof dry weight by an independent method. Even if anthro-volume (no reducing coefficient should be applied to thispometric estimates of body composition are correctedpart of the body weight). Therefore, the estimate of V

at actual weight consists of V at dry weight, estimated by for volume disturbances, the deficits of the estimates


discussed for subjects without volume disturbances will not taken into account by the anthropometric formulas,including the normal variation in the hydration of FFMstill be operative.

The current formulas are inaccurate as means of ana- and some conditions that determine body composition,such as nutrition, exercise, and the presence of wastinglyzing body composition in individual patients on perito-

neal dialysis. This lack of accuracy is the reason why disease. Systematic errors can result from obesity, proba-bly because of nonlinearity in the relation between bodythese formulas are not used routinely to estimate BF

and FFM in clinical practice. The estimates of FFM and water and body weight as weight increases, and hydrationdisorders. Therefore, anthropometric formulas can onlyBF are linked to those of body water, as shown earlier.

Consequently, the magnitude of potential errors in esti- provide approximations of body composition in perito-neal dialysis patients and their estimates should be com-mating body water by these formulas is the same as the

magnitude of their potential errors in estimating FFM pared to estimates from other methods.and BF. Yet, these formulas are routinely used for esti-mating V for urea kinetic studies in peritoneal dialysis. ACKNOWLEDGMENTIt is conceivable that the inaccurate estimates of Kt/V This work was supported by the New Mexico VA Health Care

System.urea have obscured real biologic effects of this parameterin studies of peritoneal dialysis subjects. For example,

Reprint requests to Antonios H. Tzamaloukas, M.D., Renal Section,Kt/V urea appears to be systematically underestimated 111C, New Mexico VA Health Care System, 1501 San Pedro, SE, Albu-

querque, NM 87108.in obese peritoneal dialysis subjects without volume ex-E-mail: [email protected] [10]. Obese peritoneal dialysis subjects reportedly

have a survival advantage [79]. The apparent paradoxREFERENCESof a good outcome despite low apparent Kt/V urea values

could potentially disappear if more accurate estimates 1. Blake PG, Sombolos K, Abraham G, et al: Lack of correlationbetween urea kinetic indices and clinical outcomes in CAPD pa-of V were introduced in the calculation of Kt/V urea.tients. Kidney Int 39:700–706, 1991

Thus, we believe, there is a need for a more accurate 2. Schmidt RJ, Dumler F, Cruz C: Indirect measures of total bodywater may confound precise assessment of peritoneal dialysis ade-clinical method of estimating body water in peritonealquacy. Perit Dial Int 13(Suppl 2):S224–S226, 1993dialysis patients. Studies of body water in peritoneal di-

3. Lang SM, Wolfram G, Gerzer R, Schiffl HL: Characterizationalysis, like the study of Johansson et al [10], with empha- of subtypes of hypertension in CAPD patients by cyclic guanosine

monophosphate. Perit Dial Int 19:143–147, 1999sis on acquiring knowledge in areas where there are out-4. Blake PG, Bargman JM, Bick J, et al: Guidelines for adequacystanding deficits, such as the effect of obesity, ethnicity,

and nutrition in peritoneal dialysis. J Am Soc Nephrol 10(Supplnutrition, physical activity, skeletal frame, and catabolic 13):S311–S321, 1999

5. Golper T, Burkart J, Blake P, et al: National Kidney Foun-illness, coupled with the development of clinical methodsdation—K/DOQI Clinical practice guidelines for peritoneal dialy-to estimate dry weight with accuracy, could lead to thesis adequacy: Update 2000. Am J Kidney Dis 37(Suppl 1):S75–

design of anthropometric formulas for estimating body S136, 20016. Tzamaloukas AH, Murata GH, Malhotra D, et al: Urea kineticcomposition that have an acceptable margin of error.

modeling in continuous peritoneal dialysis patients. Effect of bodyBIA offers an alternative way for analyzing body compo-composition on the methods computing urea volume of distribu-

sition for clinical purposes. This method is thought to be tion. ASAIO J 39:M359–M362, 19937. Dumler F, Cruz C: The method used for volume estimation sig-accurate in analyzing body composition in normal sub-

nificantly influences KprT/V results in peritoneal dialysis patients.jects [80–82]. However, its accuracy is diminished in dial- Adv Perit Dial 11:88–92, 1995ysis patients, with margins of error rivaling or exceeding 8. Flanigan MJ, Frankenfield DL, Prowant BF, et al: Nutrition

markers during peritoneal dialysis: Data from the 1998 peritonealthose of the anthropometric formulas [20, 27, 36, 43, 44,dialysis core indicators study. Perit Dial Int 21:345–354, 200158, 83-86]. Therefore, further systematic studies of BIA 9. Tzamaloukas AH, Murata GH: Estimating urea volume in ampu-

as a method analyzing body composition in dialysis pa- tees on peritoneal dialysis by modified anthropometric formulas.Adv Perit Dial 12:61–65, 1996tients are needed to identify and correct the potential

10. Johansson AC, Samuelsson O, Attman PO, et al: Limitations insources of error. If such studies are successful, BIA will anthropometric calculations of total body water in patients onsubstitute for anthropometric formulas as the standard peritoneal dialysis. J Am Soc Nephrol 12:568–573, 2001

11. Watson PE, Watson ID, Butt PD: Total body water volumes formethod for estimating both body water and body compo-adult males and females estimated from simple anthropometricsition in dialysis patients. measurements. Am J Clin Nutr 33:27–39, 1980

12. Hume R, Weyers E: Relationship between total body water andsurface area in normal and obese subjects. J Clin Pathol 24:234–

CONCLUSION 238, 197113. Chumlea WC, Guo SS, Zeller CM, et al: Total body water refer-The current anthropometric formulas estimating body ence values and prediction equations for adults. Kidney Int 59:

water have the potential of introducing unpredictable or 2250–2258, 200114. Lee SW, Song JH, Kim GA, et al: Assessment of total body waterpredictable (systematic) errors in the estimates of body

from anthropometry-based equations using bioelectrical imped-water, FFM, and BF in peritoneal dialysis patients. Un- ance as reference in Korean control and haemodialysis patients.

Nephrol Dial Transplant 16:91–97, 2001predictable errors can result from the presence of factors


15. Chertow GM, Lazarus JM, Lew NL, et al: Development of a of total body water in hemodialysis patients. Kidney Int 46:1438–1442, 1994population-specific regression equation to estimate total body

39. Chertow GM, Lowrie EG, Wilmore DW, et al: Nutritional assess-water in hemodialysis patients. Kidney Int 51:1578–1582, 1997ment with bioelectrical impedance analysis in maintenance hemo-16. Mellits ED, Cheek DB: The assessment of body water and fatnessdialysis patients. J Am Soc Nephrol 6:75–81, 1995from infancy to adulthood. Monographs Soc Res Child Dev (Serial

40. Kushner RF, de Vries PM, Gudivaka R: Use of bioelectrical140) 35:12–26, 1970impedance measurements in the clinical management of patients17. Wang Z, Deuremberg P, Wang W, et al: Hydration of fat-freeundergoing dialysis. Am J Clin Nutr 64(Suppl 3):503S–509S, 1996mass: New physiological modeling approach. Am J Physiol 276:

41. Chertow MG, Lazarus JM, Lew LN, et al: Bioimpedance normsE995–E1003, 1999for the hemodialysis population. Kidney Int 52:1617–1621, 199718. Wang Z, Deuremberg P, Wang W, et al: Hydration of fat-free

42. Zhu F, Schneditz D, Levin NW: Sum of segmental bioimpedancebody mass: Review and critique of a classic body-composition con-analysis during ultrafiltration and hemodialysis reduces sensitivitystant. Am J Clin Nutr 69:833–841, 1999to changes in body position. Kidney Int 56:692–699, 199919. Ramdeen G, Tzamaloukas AH, Malhotra D, et al: Estimates of

43. Cooper BA, Aslani A, Ryan M, et al: Comparing different meth-interdialytic sodium and water intake based on the balance princi-ods of assessing body composition in end-stage renal failure. Kid-ple. Differences between nondiabetic and diabetic subjects on he-ney Int 58:408–416, 2000modialysis. ASAIO J 44:812–817, 1998

44. Kooman JP, Cox-Reijven PLM, van der Sande FM, et al: Assess-20. de Fijter CWH, de Fijter MM, Oe LP, et al: The impact ofment of body composition in ESRF [letter]. Kidney Int 59:692–hydration status on the assessment of lean body mass by body699, 2001electrical impedance in dialysis patients. Adv Perit Dial 9:101–

45. Schmidt R, Dumler F, Cruz C, et al: Improved nutritional follow-104, 1993up of peritoneal dialysis patients with bioelectrical impedance. Adv21. James WPT, Bailes J, Davies HL, Dauncey MJ: Elevated meta-Perit Dial 8:157–159, 1992bolic rates in obesity. Lancet 1:1122–1125, 1978

46. Bergia R, Bellini ME, Valenti M, et al: Longitudinal assessment22. Arkouche W, Fouque D, Pachiaudi C, et al: Total body composi-of body composition in continuous ambulatory peritoneal dialysistion in chronic peritoneal dialysis patients. J Am Soc Nephrol 8:using bioelectrical impedance and anthropometric measurements.1906–1914, 1997Perit Dial Int 13(Suppl 2):S512–S514, 199323. de Fijter WM, de Fijter CWH, Oe PL, et al: Assessment of

47. Schmidt RJ, Dumler F: Bioelectrical impedance analysis: A prom-total body water and lean body mass from anthropometric Watsonising predictive tool for nutritional assessment in continuous ambu-formula, creatinine kinetics and body electrical impedance com-latory peritoneal dialysis patients. Perit Dial Int 13:250–255, 1993pared with antipyrine kinetics in peritoneal dialysis patients.

48. Bhatla B, Moore H, Emerson P, et al: Lean body mass estimationNephrol Dial Transplant 12:151–156, 1997by creatinine kinetics, bioimpedance and dual energy x-ray absorp-24. Woodrow G, Oldroyd B, Turney JH, et al: Measurement of totaltiometery in patients on continuous ambulatory peritoneal dialysis.body water and urea kinetic modelling in peritoneal dialysis. ClinASAIO J 41:M442–M446, 1995Nephrol 47:52–57, 1997

49. Briganti M, Emiliani G, Montanari A, et al: Longitudinal assess-25. Dahl NV, Foote EF, Kapoian T, et al: Measuring total body water ment of body composition in CAPD patients using bioelectric im-in peritoneal dialysis patients using an ethanol dilution technique. pedance analysis. A comparison to hemodialysis patients. ASAIOKidney Int 56:2297–2303, 1999 J 41:M725–M727, 1995

26. Panzetta G, Guerra U, D’Angelo A, et al: Body composition and 50. Caillette A, Labeeuw M, Zech P: Total body water determinationnutritional status in patients on continuous ambulatory peritoneal using tetrapolar impedancemetry in CAPD patients. Nephrol Dialdialysis (CAPD). Clin Nephrol 23:18–25, 1985 Transplant 9:585, 1995

27. Kong CH, Thompson CM, Lewis CA, et al: Determination of total 51. Smye SW, Bradbury M, Brockenbank T: Total body water deter-body water in uraemic patients by bioelectrical impedance. Nephrol mination using tetrapolar impedancemetry in CAPD patients.Dial Transplant 8:716–719, 1993 Nephrol Dial Transplant 9:585–586, 1995

28. Wong KC, Yiong DW, Kerr PG, et al: Kt/V in CAPD by different 52. Edefonti A, Carcano A, Damiani B, et al: Changes in body com-estimations of V. Kidney Int 48:563–569, 1995 position assessed by bioimpedance analysis in the first 6 months

29. Durnin JVCA, Womersley J: Body fat assessed from body dens- of chronic peritoneal dialysis. Adv Perit Dial 13:267–270, 1997ity and its estimation from skinfold thickness. Measurements in 53. Domoto ST, Weindel ME: Bioimpedance analysis of fluid com-481 men and women aged from 16 to 72 years. Br J Nutr 32:77– partments in female CAPD patients. Adv Perit Dial 14:220–222,97, 1974 1998

30. Jager KJ, Mercus MP, Husman RM, et al: Nutritional status over 54. Dumler F, Falla P, Butler R, et al: Impact of peritoneal dialysistime in hemodialysis and peritoneal dialysis. J Am Soc Nephrol modality and acidosis on nutritional status in peritoneal dialysis12:1272–1279, 2001 patients. Adv Perit Dial 14:205–208, 1998

31. Nyboer J, Sedensky JA: Bioelectrical impedance during renal 55. Jones CH, Smye SW, Newstead CG, et al: Extracellular fluiddialysis. Proc Clin Dial Transplant Forum 4:214–219, 1974 volume determines by bioelectric impedance and serum albumin

32. Bohm D, Odaischi M, Beyerlein C, Overbeck W: Total body in CAPD patients. Nephrol Dial Transplant 13:393–397, 1998water: Changes during dialysis estimated by bioimpedance analysis. 56. Passadakis P, Sud K, Dutta A, et al: Bioelectrical impedanceInfusiontherapie 17 (Suppl 3):75–78, 1990 analysis in the evaluation of the nutritional status of continu-

33. Kurtin PS, Shapiro AC, Tomita H, Raizman D: Volume status ous ambulatory peritoneal dialysis patients. Adv Perit Dial 15:147–and body composition of chronic dialysis patients: Utility of bio- 152, 1999electrical impedance plethysmography. Am J Nephrol 10:363– 57. Song JH, Lee SW, Kim GA, Kim MK: Measurement of fluid shift in367, 1990 CAPD patients using segmental bioelectrical impedance analysis.

34. de Lorenzo A, Barra PF, Sasso GF: Body impedance measure- Perit Dial Int 19:386–390, 1999ments during dialysis. Eur J Clin Nutr 45:321–325, 1991 58. Than N, Woodrow G, Oldroyd B, et al: Effect of peritoneal fluid

35. Dumler F, Schmidt RJ, Kilates C, et al: Use of bioelectrical on whole body and segmental multiple frequency bioelectricalimpedance for the nutritional follow-up of chronic hemodialysis impedance in patients on peritoneal dialysis. Eur J Clin Nutr 54:patients. Miner Electrolyte Metab 18:284–287, 1992 450–451, 2000

36. Kouw PM, Olthof CG, ter Wee PM, et al: Assessment of post- 59. Zhu F, Schneditz D, Kaufman AM, Levin NW: Estimation ofdialysis dry weight: An application of the conductivity measure- body fluid changes during peritoneal dialysis by segmental bi-ment method. Kidney Int 41:440–444, 1992 oimpedance analysis. Kidney Int 57:299–306, 2000

37. Pastan S, Gassensmith C: Total body water measured by bioelec- 60. Edefonti A, Picca M, Damiani B, et al: Prevalence of malnutritiontrical impedance in patients after hemodialysis. Comparison with assessed by bioimpedance analysis and anthropometry in childrenurea kinetics. ASAIO J 38:M186–M189, 1992 on peritoneal dialysis. Perit Dial Int 21:172–179, 2001

61. Nielsen PK, Ladefoged J, Olgaard K: Lean body mass by dual38. Ho LT, Kushner RF, Schoeller DA, et al: Bioimpedance analysis


energy x-ray absorptiometry (DEXA) and by urine and dialysate patients on continuous ambulatory peritoneal dialysis, in Continu-ous Ambulatory Peritoneal Dialysis, edited by Legrain M, Amster-creatinine recovery in CAPD and predialysis patients compared

to normal subjects. Adv Perit Dial 10:99–103, 1994 dam, Excerpta Medica, 1980, pp 263–26775. Afthentopoulos IE, Shetty A, Oreopoulos DG: Hypotension62. Stall S, Ginsberg NS, Lynn RI, Zabetakis PM: Bioelectrical

impedance analysis and dual energy x-ray absorptiometry to moni- in CAPD: Role of volume and sodium depletion. Adv Perit Dial11:73–77, 1995tor nutritional status. Perit Dial Int 15 (Suppl 5):S59–S62, 1995

63. Woodrow G, Oldroyd B, Turney JH, et al: Four-component 76. Tzamaloukas AH: Effect of edema on urea kinetic studies inperitoneal dialysis. Perit Dial Int 14:398–401, 1994model of body composition in chronic renal failure comprising

dual-energy x-ray absorptiometry and measurement of total body 77. Tzamaloukas AH, Murata GH, Dimitriadis A, et al: Fractionalurea clearance in continuous peritoneal dialysis: Effects of volumewater by deuterium oxide dilution. Clin Sci 91:763–769, 1996

64. Lukaski HC: Validation of body composition assessment tech- disturbances. Nephron 74:567–571, 199678. Tzamaloukas AH, Murata GH, Malhotra D, et al: Normaliza-niques in the dialysis population. ASAIO J 43:251–255, 1997

65. Keshaviah PR, Nolph KD, Moore HL, et al: Lean body mass tion of clearances in peritoneal dialysis using a formula for bodywater derived from an end-stage renal disease population. Peritestimation by creatinine kinetics. J Am Soc Nephrol 4:1475–1485,

1994 Dial Int 20:60–64, 200079. Johnson DW, Herzig KA, Purdie DM, et al: Is obesity a favorable66. Hewitt MJ, Going SB, Williams DP, Lohman TG: Hydration of

the fat free body mass in children and adults: Implications for body prognostic factor in peritoneal dialysis patients? Perit Dial Int20:715–721, 2000composition assessment. Am J Physiol 265:E88–E95, 1993

67. Wang Z, Deuremberg P, Heymsfield SB: Cellular-level body com- 80. Hoffer EC, Meador CK, Simpson DC: Correlation of whole bodyimpedance with total body water volume. J Appl Physiol 27:531–position model. A new approach to studying fat-free mass hydra-

tion. Ann N Y Acad Sci 904:306–311, 2000 534, 196981. Kushner RF, Schoeller DA: Estimation of total body water by68. Metropolitan height and weight tables: Stat Bull Metrop Life

Insur Co 64:3, 1983 bioelectrical impedance analysis. Am J Clin Nutr 44:417–424, 198682. van Marken Lichtenbelt WD, Westerterp KR, Wouters L, Lui-69. Johansson AC, Samuelsson O, Haraldsson B, et al: Body compo-

sition in patients treated with peritoneal dialysis. Nephrol Dial jendijk SCM: Validation of bioelectric-impedance measurementsas a method to estimate body-water compartments. Am J Clin NutrTransplant 13:1511–1517, 1998

70. Jolly S, Chattatalsingh C, Bargman J, et al: Excessive weight 60:159–166, 199483. Rallison LR, Kushner RF, Penn D, Schoeller DA: Errors ingain during peritoneal dialysis. Int J Artif Organs 24:197–202, 2001

71. James WPT, Reeds PJ: Nutrient partitioning, in Handbook of estimating peritoneal fluid by bioelectrical impedance analysis andtotal body electrical conductivity. J Am Coll Nutr 12:66–72, 1993Obesity, edited by Bray GA, Bouchard C, James WPT, New York,

Marcel Dekker, 1998, pp 555–571 84. Bergia R, Bellini ME, Dionisio P, et al: Is bioelectrical impedanceactually useful for nutritional assessment in patients on CAPD?72. Coles GA: Have we underestimated the importance of fluid bal-

ance for the survival of PD patients? Perit Dial Int 17:321–326, Perit Dial Int 14:290–291, 199485. Woodrow G, Oldroyd B, Turney JH, et al: Measurement of total1997

73. Tzamaloukas AH, Saddler MC, Murata GH, et al: Symptomatic body water by bioelectrical impedance in chronic renal failure.Eur J Clin Nutr 50:676–681, 1996fluid retention in patients on continuous peritoneal dialysis. J Am

Soc Nephrol 6:198–206, 1995 86. Di Iorio BR, Tarraciano V, Bellizzi V: Bioelectrical impedancemeasurement: Errors and artifacts. J Ren Nutr 9:192–197, 199974. Marquez-Julio A, Dombros N, Osmond D, et al: Hypotension in

estimates of body water, fat-free mass, and body fat in patients on peritoneal dialysis by...

Documents