- Email: [email protected]
Body fluid retention and body weight change in anorexia nervosa patients during refeeding
Clinical Nutrition 29 (2010) 749e755
Contents lists available at ScienceDirect
Clinical Nutrition journal homepage: http://www.elsevier.com/locate/clnu
Original Article
Body fluid retention and body weight change in anorexia nervosa patients during refeeding Daniel Rigaud a, *, Alain Boulier b, Isabelle Tallonneau a, Marie Claude Brindisi a, Raymond Rozen b a b
Service d’Endocrinologie et Nutrition, CHU Le Bocage, B.P. 77 908, 21079 Dijon, France Service d’explorations fonctionnellles, CHU Bichat, 75018 Paris, France
a r t i c l e i n f o
s u m m a r y
Article history: Received 16 December 2009 Accepted 28 May 2010
Background & aims: Body weight gain is an important goal in anorexia nervosa (AN) patients, but inflation in body fluids could artificially increase body weight during refeeding. Methods: 42 malnourished adult AN patients were refed using a normal-sodium diet, then 176 other malnourished adult AN patients received a refeeding low-sodium diet (BMI of the 218 patients: 13.4 1.9 kg/m2). Sodium balance, body composition by a 2-electrode impedance method (BIA, for assessment of total and extracellular water, fat-free mass, FFM), resting energy expenditure and energy intake were calculated. Results: In the patients on normal-sodium diet, body weight, and total and extracellular water gains were higher than those of the low-sodium diet patients (P < 0.01). Edema occurred more often in the former group (21% vs 6%; P < 0.05). In almost all patients, BMI reached a plateau around 15e16 kg/m2, then increased again. During this plateau, an increase in intracellular water and in “active FFM” was observed with BIA, together with a similar decrease in extracellular water. Conclusion: In AN patients, who are always afraid of gaining too much weight, in regard to their food intake, it will be useful to give a low-sodium diet until a 15e16 kg/m2 BMI. This should be integrated into the cognitive behavioral therapy. Ó 2010 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Keywords: Anorexia nervosa Body composition Eating disorder Malnutrition Tube feeding
1. Introduction In anorexia nervosa (AN), body weight is a key factor for medical decision. It has been described as one of the predictive factors for relapse.1e4 It has been included in different therapeutic goals and cognitive behavioral therapies.1,2,5 Reaching a normal BMI, for example a BMI at or above 18.5 kg/m2 in adult patients, seems a determinant for complete recovery. In AN patients, hospitalization is often decided because of body weight loss, as assessment of malnutrition. For some medical teams, body weight (BW) remains the primary goal of the contract signed by the hospitalized malnourished AN patients with the medical staff.6 In this “bodyweight contract”, some teams give to the patient’s authorization for familial visits or outdoor permissions according to body weight gain. In almost all the teams, the decision concerning the discharge from the hospital is determined by the BW.1,2,6 Lastly, when an AN patient is suspected to lie, because she doesn’t gain body weight despite high food and energy intake, it should be very important to * Corresponding author. Service d’Endocrinologie et Nutrition, CHU Le Bocage, 21000 Dijon, France. Tél.: þ33 3 80 29 34 52; fax: þ33 3 80 29 35 19. E-mail address: [email protected] (D. Rigaud).
know whether this lack of weight gain has to be interpreted as an absence of active mass gain, or as a gain in active mass associated with a loss in body water. Inversely, in the case of weight gain, the medical staff should know if this gain represents true active body mass, or only gain in fluids (and sodium) in extracellular compartments. Indeed, it is well known that, in several patients, malnutrition is associated with abnormalities of sodium-potassium pumps, inducing an increase in sodium and water retention.7 Legs or truncal edema are well described in malnourished non AN and AN patients,7,8 even if large edema are more frequent in Kwashiorkor than in AN malnourished patients. It is self evident that, in malnourished patients, body weight gain is partly composed of intracellular fluids in fat-free mass cells. But a gain in extracellular water could also be observed, in particular on free-sodium diet (10e12 g NaCl per day). Very few studies including only a small number of malnourished AN patients submitted to refeeding are published.8e11 Since we had observed that some patients, during refeeding, gained more body weight than expected from their energy balance, we decided to perform the present study to describe the changes in body weight, body masses and body fluids in 218 malnourished AN
0261-5614/$ e see front matter Ó 2010 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved. doi:10.1016/j.clnu.2010.05.007
750
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755
patients hospitalized for refeeding, either refed using a low-sodium or a normal-sodium diet. 2. Patients and methods 2.1. Patients In this study, the chart of all the consecutive malnourished adult AN women who were hospitalized for renutrition during the last 4 yrs in our Nutrition unit for at least 2 months was analyzed. The characteristics of these 218 malnourished AN patients were seen in Table 1. During the same period (4 yrs), 284 AN patients were hospitalized: 66 were not included in the analysis, because having a BMI >16 kg/(m)2 (n ¼ 8), being no compliant, refusing or stopping tube feeding, engaging frequent vomiting or purging episodes during stay, having too short stay (<2 months). In the first 42 patients, refeeding was performed using a combination of normal-sodium diet and nasogastric-tube feeding. In all, resting energy expenditure and body composition by bioimpedance were measured, as described before.12,13 Twenty four patients suffered from the restrictive subtype (RAN, 57%) and 18 others from the binge/purging subtype (BPAN) of the disease (DSM IV criteria). During their refeeding, there was no restriction in water or sodium intake from meals or snacks (normal-sodium diet group). The enteral mixture for tube feeding contained 0.6 g NaCl per liter (Sondalis isoÒ: 1000 mL ¼ 1000 kcal). Because, in our initial experience, several malnourished adult AN patients developed edemas or gained “too much” body weight, in regard of their low energy intake, we decided to use low-sodium diet in such patients. In the further 176 malnourished AN patients, was prescribed a low-sodium diet (4e5 g NaCl/day per 2000 kcal/ day). There were 92 RAN patients (56%) and 84 BPAN (44%). All had, as the previous 42 patients, meals and tube feeding for refeeding. The low-sodium diet was changed to a normal-sodium diet (10e12 g NaCl/day) as soon as the sodium urinary output has clearly increased. Among these 176 patients, 21 patients did not accept tube feeding anymore after a mean duration of 23 5 days of enteral nutrition. The criteria for non inclusion were diuretics or laxative misuse, corticotherapy, low compliance, or purging behavior. These AN patients were compared with 116 healthy women without eating disorder (Table 1). 2.2. Methods The following measurements were made every week in each patient:
2.2.1. Body composition 2.2.1.1. Anthropometry. Body height was measured at admission at the nearest lower 0.5 cm. Two times a week, body weight was measured at the nearest 100 g (reproducibility: 50 g), and BMI calculated. Skinfold thickness was determined at the 4 classical area (triceps, biceps, subscapular, suprailiac) using metallic Harpenden calliper and then body FM and FFM were calculated as follow: body density ¼ 1.1599[0.0717 (log S)], where S is the sum of the 4 skinfold.14 2.2.1.2. Bioelectrical impedance. It was used to measure body composition, as previously described.15 We obtained, using this two-frequencies analysis (impédancemètre Innerscan 2-fr BIA Tanita corp, France), the extracellular water (ECW) and body water (TW), then intracellular water (ICW) and fat-free mass (FFM), then active FFM, then fat mass (FM) from body weight. In this study, a double-frequency bioimpedance analyser15 with subcutaneous electrodes (using acupuncture needle with strict asepsis) was used, in order to obtain values for total body water (high frequency, 1 MHz) and extracellular water (low frequency, 50 kHz). One electrode was introduced under the skin of the right hand and the other under the skin of the left foot. Briefly, calculations were as following: Z2 ¼ R2 þ Xfc, Xfc¼(Z) sin q, R¼(Z) cos q; total body water (V): r H/S and V ¼ S H (R ¼ resistance, Xfc ¼ reactance, r ¼ resistivity, H ¼ height; FFM ¼ V/0.73; extracellular body water: V at 50 kHz; total body water ¼ V at 1 MHz, intracellular body water: totaleextracellular water; active fat-free mass ¼ FFMextracellular water; fat mass: body weight (measured just before BIA) minus fat-free mass. 2.2.2. Other measurements 2.2.2.1. Blood and urine analyses. Every week, total 48-h urinary output of sodium, potassium, urea and creatinine were measured; albumin, transthyrétine and hematocrit were obtained every two weeks. Energy intake was calculated from values of a food table (BilnutÒ, 1996), as previously described.12,13 We checked the remaining mixtures (tube feeding) and the remaining food after meals. In 39 patients on normal-sodium and 168 on low-sodium diet, resting energy expenditure (REE) was measured by indirect calorimetry.13 In the others, it was calculated from Harris-Benedict formulas. From these values, energy balance was calculated as follow: energy intake (from meals and tube feeding) minus energy needs. Energy needs during hospital stay were calculated as previously determined, i.e. resting energy expenditure 1.5 in patients having not excessive exercise and 1.7 in hyperactive patients. Theoretical mass retention (gain) was calculated on the
Table 1 Body composition at hospital’s admission in the 42 normal-sodium and the 176 low-sodium AN patients and in the 116 normal-weight healthy controls.
Age (years) Women (%) Body weight (kg) Body mass index (kg/(m)2) Total body water (kg) Total body water (%) Extracellular (EC, kg) Intracellular (IC, kg) EC/IC Fat-free mass (kg) Fat mass (kg) Active fat-free mass (kg) Fat-free mass (%) Fat mass (%)
Anorexia nervosa (n ¼ 42)
Anorexia nervosa (n ¼ 176)
Controls (n ¼ 116)
Expected body composition at 36 kga
22.1 4.2 97% 36.6 4.3 ** 13.8 1.7 ** 28.7 5.3 ** 78.6 3.0 ** 15.6 3.1 13.1 2.1 ** 1.19 0.16 ** 34.7 4.6 ** 1.9 1.7 ** 19.1 3.5 ** 94.8 5.1 ** 5.2 5.1 **
23.3 5.1 98% 36.0 3.8 ** 13.2 1.2 ** 28.8 4.8 ** 80.0 3.3 ** 15.3 2.8 13.5 2.5 ** 1.13 0.14 ** 34.1 4.6 ** 2.3 2.2 ** 18.8 3.1 ** 94.7 5.4 ** 5.3 5.1**
27.2 8.8 110 (95%) 57.6 9.3 22.3 2.6 35.6 4.7 61.8 3.4 16.1 2.1 19.5 2.7 0.82 0.05 44.6 7.1 13.0 4.3 31.2 4.9 77.4 4.1 22.6 4.1
e 36 4.4 13.4 1.9 23 2.4 63.9 3.6 10 1.4 13 1.8 0.77 0.06 29 2.4 7 1.7 19 2.2 80.6 4.5 19.4 3.1
**P < 0.005 vs controls; NS between AN patients. a As expected from body composition of the 116 control subjects at a body weight of 36 kg. Active fat-free mass: fat-free mass minus extracellular water.
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755
2.2.3. Statistical analyses Mean and standard deviation were given. From two-frequencies bioelectrical impedance (BIA), were measured total body water, extracellular water and fat-free mass. The active fat-free mass (CFFM) derived from fat-free mass, i.e. fat-free mass minus extracellular water (FFM ECW), as both obtained from BIA. From 48-h urinary sodium output and 48-h sodium intake (by meals and tube feeding) obtained during 2 consecutive days a week, we determined the total water body gain, using the following formula: 140 mmol Na þ gain, i.e. 9 g NaCl gain, corresponded to a gain of 1 L, i.e. one kilo of body weight. Creatinine urinary output was calculated as the mean of two successive 24-h urinary outputs. The results were analyzed according to BMI, with classes of BMI for each point of BMI (ex: 12e12.9, 13e13.9.). Student t test was used for comparisons between groups. Non parametric Wilcoxon tests were performed to compare two periods of time in our patients. 3. Results Tolerance of tube feeding was good: no severe side effect, no refeeding syndrome, no metabolic problem, 12 sinusitis (6%). All patients had normal hematocrit, transthyretin and albumin levels during all the study period (2 months). After 2 months, a decrease in albumin (from 42 to 38 g/L, P < 0.01) and in hematocrit level (from 41 to 38%, P < 0.01) was observed. The first 42 malnourished AN patients were refed using a free normal-sodium diet (3 meals and 1 or 2 snacks) and tube feeding, in order to complement energy input. The increase in body weight was 1.42 0.23 kg after one week of refeeding, despite an energy intake around their energy needs (1638 234 kcal/day). In 9 out of the 42 patients (21%), all having a BMI<15 kg/(m)2, indisputable legs edema were observed during renutrition. This was associated with a mean body weight gain of 1.34 0.31 kg per week during the first 3 weeks of refeeding. In two of them, having the BPAN subtype, the edema were large. In 8 of these 9 patients, the edema disappeared when their BMI reached a value between 15 and 17 kg/(m)2. In the last one (BPAN patient), edema persisted until the discharge from the hospital. The further 176 malnourished AN patients were refed using a free (as they could eat) but low-sodium diet including 3 meals and 1 or 2 snacks. Renutrition also included tube feeding. Their body weight increase was 0.73 0.18 kg after one week of refeeding, with an energy intake around their energy needs (1583 207 kcal/ day). The low-sodium diet was given until the urinary sodium output clearly increased. This increase in urinary output often occurred after 4e6 weeks of refeeding. In 11 patients (6%; BPAN: n ¼ 8), all having a BMI<15 kg/(m)2, light edema of the legs were observed during renutrition. Edema were less frequent than in the normal-sodium diet group (chi square; P < 0.05). The patients on low-sodium diet gained 9.3 0.9 kg of body weight during the 2-month refeeding. BMI increased from 12.6 1.2 to 17.6 1.2 kg/(m)2. After one month refeeding, energy input reached 2448 162 kcal/day (including 886 115 kcal by tube feeding) in the normal-sodium diet group and 2512 171 kcal/day (including 924 121 kcal by tube feeding) in the low-sodium diet group. After 2 months, energy input reached
2684 186 kcal/day (812 109 kcal by tube feeding) in the normalsodium diet group and 2647 192 kcal/day (862 92 kcal by tube feeding) in the low-sodium diet group. There was no difference in these energy input between the low- and the normal-sodium group (ANOVA, F ¼ 0.87). Until a mean BMI of 15e16 kg/(m)2, the mean BMI of the 2 groups (low- and normal-sodium diets) increased linearly. Around a BMI value of 15.2 0.4 kg/(m)2, the body weight remained stable during 7e10 days. This plateau occurred at the threshold of BMI 15e16 kg/(m)2 in 192 patients (88%). This stability of body weight contrasted with the absence of any decrease in energy intake, which was not significantly different from those of the previous and the next weeks. Before it reached this 15e16 kg/(m)2 value, BMI of the patients in the normal-sodium diet increased more than BMI of the low-sodium diet patients (P < 0.05). This occurred despite similar resting energy expenditure and similar energy input by meals and tube feeding in the 2 groups (see above). 3.1. Sodium urinary output The course of sodium urinary output was seen in Fig. 1. In the 34 normal-sodium diet patients for whom we had sodium daily balance, the sodium balance was positive until the BMI reached a value of 15e16 kg/(m)2, then it tended to be negative. In the 71 patients on low-sodium diet for whom we had sodium daily balance, the sodium balance was less positive (difference between groups, P < 0.01). In the two groups, mean sodium balance became moderately negative, when BMI reached a value of 15e16 kg/(m)2. 3.2. Body composition with the two-frequency bioelectrical impedance analysis (2-fr BIA) Table 1 gives the results of BIA in the AN patients, compared with the healthy subjects. The AN patients had higher total body water and higher extracellular water (P < 0.01). Total body water represented 78.4 3.2% of body weight in the AN patients and 63.9 3.6% in the controls (P < 0.01). The extra-/intracellular water ratio was higher in AN patients than in controls. Compared with values extrapolated from control values at a 36 kg-body weight, the AN patients had lower fat mass (P < 0.01), and higher fat-free mass than these shadow controls (P < 0.01), but similar active FFM (NS).
30
25
20
15 mmol/24 hr
basis of 7.5 kcal/g body weight storage. This allowed us to expect the body weight gain by week, and to conclude whether or not the body weight gain was in accord with the energy-excess. The protocol of this study was approved by the ethical committee of Paris 7 university. Indirect calorimetry (for REE) and bioelectrical impedance (for body composition) were performed in all patients admitted in our unit in this period, even those who did not participate in the present study. Each patient gave informed consent to participate.
751
10
*
*
*
*
11-12
12-13
13-14
14-15
5
0
-5
-10 15-16
16-17
17-18 2
Normal sodium diet
low sodium diet
BMI (kg/m )
Fig. 1. Evolution of daily sodium balance in 34 patients refed by normal-sodium diet and in 71 patients on low-sodium diet. Normal-sodium diet: sodium intake 10.3 2.3 g NaCl/day; low-sodium diet: 4.3 1.2 g NaCl/day (including tube feeding: 600 mg/L [1000 kcal]). The fact that sodium balance was going to be negative from P < 0.05. a BMI value of 15e16 kg/m2 must be noted.
752
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755
Table 2 Changes in body composition during refeeding in the 176 anorexia nervosa patients refed with a low-sodium diet, according to the type of the disease.
Body weight (kg) BMI (kg/(m)2) Total body water (kg) Extracellular W (kg) Intracellular W (kg) Extra-/intra water ratio Fat-free mass (kg) Fat mass (kg) Active FFM (kg)
RAN (n ¼ 92) admission
RAN (n ¼ 90) 2 months
BPAN (84) admission
BPAN (n ¼ 81) 2 months
35.2 4.1 13.1 1.2 26.9 2.1 14.5 1.3 12.4 1.0 1.17 0.11 33.4 1.6 1.8 0.5 18.9 1.1
44.4 2.1 16.8 1.4 29.7 1.9 13.3 1.1 16.5 1.5 0.80 0.08 $ 39.1 1.7 5.3 0.6 25.8 1.5
38.4 3.7 * 14.3 0.7 * 24.5 1.8 12.3 1.2 * 12.2 1.5 1.01 0.12 * 35.1 2.8 3.7 1.5 * 22.8 1.8 *
48.4 2.5 17.9 1.8 28.1 2.0 12.6 1.2 15.5 1.5 0.81 0.10 42.5 2.0 7.0 0.8 29.9 1.3
RAN: restrictive subtype, BPAN: binge-eating/purging subtype. W: water. Results were obtained using two-frequencies bioelectrical impedance analysis at admission and after a 2-month refeeding (mean of 57 6 days). A low-sodium diet (<4 g NaCl/day) was given. As shown, despite a body weight gain of 8e10 kg, extracellular water did not increase. *P < 0.02 versus RAN (patients suffering from self-induced vomiting). In the 2 groups, the values obtained at 2 months were significantly higher than those obtained before the start of refeeding (P < 0.001), except for extracellular water and EW/IW ratio.
This suggested that the water content of FFM was too high in AN patients. Table 2 gives the results of BIA in the AN patients, according to the type of the disease. The patients having the binge-eating/ purging subtype (BPAN) had higher body weight (and BMI), higher FFM and higher FM than the RAN patients (P < 0.01). For this higher body weight, the BPAN patients had lower total body water and lower extracellular water (non significant) than the RAN patients. As shown in Table 2, the extra-/intracellular water ratio, which was elevated before refeeding, decreased with refeeding, while the patients had reached a BMI higher than 15 kg/(m)2. After 2-month refeeding, only two variables were not increased: extracellular water and extra-/intracellular ratio. Body composition changes differed between the normalsodium and the low-sodium diets (Table 3). On normal-sodium diet, the patients gained more body weight, more FFM than the patients on low-sodium diet (P < 0.05). As shown in Fig. 2, the increase in body weight during refeeding (P < 0.001) was in part due to an increase in extracellular fluids until the fourth week (P < 0.02). The curve of active FFM and intracellular water were linear, although that of extracellular water was omega shaped (maximal values between week 3 and week 5 in the two groups). The active FFM (FFM minus the extracellular water) and the intracellular water tended to increase at a constant rate, while extracellular water did not. Similarly, between the 4th and the 6th weeks, while energy input did not decrease, body weight and FFM did not increase anymore, for one to two weeks. This corresponded to a BMI value between 15 and 16 kg/(m)2. During the same period, fat mass, ICW and active FFM increased
linearly in the 2 diet groups (Fig. 2). By contrast, extracellular water showed different evolution in the two diet groups: ECW did not change in the low-sodium group, while it increased then decreased in the normal-sodium diet group. 3.3. Relationship between skinfold thickness and 2-fr-bioelectrical impedance analysis As determined from the four skinfold thickness, FM increased more using skinfold thickness measurement than using BIA. The gain of FM after one month refeeding was þ1.8 0.2 kg using skinfold thickness versus þ1.2 0.3 kg using BIA in the normalsodium group and þ1.8 0.2 kg (skinfold thickness) vs þ1.2 0.2 kg using BIA in the low-sodium diet group (P < 0.05 for each). There was a good correlation for FM between the values obtained from skinfold thickness and BIA (r ¼ 0.684, P < 0.001). But BIA systematically underestimated FM. For example, 18 out of our 218 patients (8%) had a BMI lower than 11 kg/(m)2: in 14 of them (77%), fat mass was negative in BIA (1.1 0.4 kg), because BIAcalculated FFM was higher than body weight. In these 14 patients, FM was þ0.9 0.2 kg using skinfold thickness. 3.4. Relationship between energy expenditure and body weight gain Considering the mean energy balance (energy intake minus daily energy expenditure) during the first month of refeeding, the expected body weight gain of the normal-sodium diet patients was
Table 3 Changes in body composition during refeeding in the normal-sodium and low-sodium diet groups, using skinfold thickness or BIA measurement.
BMI (normal-sodium diet) BMI (low-sodium diet) Weight (normal-sodium diet) Weight (low-sodium diet) FFM (NSD) skinfold FFM (low SD) skinfold FMM (NSD) BIA FMM (low SD) BIA Fat mass (NSD) skinfold Fat mass (low SD) skinfold Fat mass (NSD) BIA Fat mass (low SD) BIA
Week 0
W2
W4
W6
W8
13.3 1.9 13.7 1.6 36.2 4.3 36.5 2.3 33.3 2.8 33.4 2.0 34.1 3.7 33.9 3.3 2.9 1.3 3.1 1.1 2.1 1.5 2.6 1.4
14.5 1.6 14.7 1.3 39.4 3.4 38.9 2.1 35.9 2.0 35.1 1.7 35.6 3.5 36.0 2.9 3.5 1.0 3.8 1.0 2.8 1.2 2.9 1.3
15.9 1.3 15.0 1.6 43.4 3.1 39.9 2.0 39.3 2.4 35.6 1.8 * 37.0 3.6 36.8 2.9 4.1 1.1 4.3 1.1 3.7 1.2 3.1 1.2
15.6 1.5 15.9 1.7 42.5 3.3 41.9 2.1 * 38.0 2.2 36.9 1.7 * 38.9 3.7 38.4 2.8 * 4.6 1.0 4.9 0.9 3.9 1.3 3.6 1.1
16.2 1.6 16.4 1.5 44.2 3.1 43.6 2.2 * 38.9 2.5 38.1 1.8 * 38.3 3.5 39.4 2.7 * 5.3 0.9 5.6 0.8 4.0 1.4 4.2 1.2
Measurements were performed each week in a mean of 36 patients on normal-sodium diet and in a mean of 164 patients on low-sodium diet. BIA: 2-fr-bioelectrical impedance analysis. Skinfold: from triceps, biceps, subscapular, suprailiac measurements. For skinfold thickness, FFM was calculated as body weight minus fat mass. For both skinfold and BIA calculation, FFM included extracellular water. * significantly different from normal-sodium diet (NSD): P < 0.05.
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755
753
28
26
24
22
kg
active FFM 20
normal-sodium diet
*
*
18
extracellular water
*
16
low-sodium diet 14
intracellular water 12 W0 active FMM NSD
W1
W2
W3
active FMM low SD
W4 ECW NSD
W5 ECW low SD
W6
W7
W8 weeks
ICW NSD
ICW low SD
Fig. 2. Evolution of active FFM, extracellular and intracellular water in the patients on normal-sodium or low-sodium diet Squares: normal-sodium water group; circles: lowsodium water group. Black: FFM, grey: extracellular water, white: intracellular water. The patients were refed both by meals and tube feeding. Active fat-free mass (FFM minus extracellular water) and intracellular water increased at the same rhythm in the 2 groups. By contrast, extracellular water increased more in the normal-sodium diet group. Notes P < 0.05 at W3, W4 and W5 between normalthat the curves of intra- and extracellular water crossed each other at W5-W6 in both groups (mean BMI: 15.34 0.28 kg/(m)2). sodium and low-sodium diet groups.
3.9 0.3 kg, while it was measured at 6.7 0.5 kg (P < 0.01 vs expected weight gain). In the low-sodium diet group, the expected weight gain was 4.0 0.4 kg and the true gain 3.8 0.3 kg (expected vs true: not significant). In the two groups, after that the BMI had reached the value of 15 kg/(m)2, the weight gain was similar as that expected. 4. Discussion In these malnourished AN patients, a BMI and FFM plateau was observed during refeeding, despite unchanged energy intake or expenditure. This plateau occurred although the patients did not reduce their energy intake. It was observed between the 4th and the 6th week of hospitalization. This plateau occurred very frequently at a 15e16 kg/(m)2 BMI. At this time, we observed a tendency for a decrease in total FFM and extracellular water, while active FFM (and intracellular water) and fat mass continued to increase. During the same period, the sodium balance, which was clearly positive during the previous weeks (in particular in the normal-sodium diet group) turned to be negative during 2e3 weeks. This was related to an increase in sodium urinary output, and not to a decrease in sodium intake. The increase in sodium loss associated with the decrease in extracellular water suggested that, at this 15e16 BMI value, the sodium and water retention related to malnutrition tended to disappear. These changes were more accurate in patients submitted to normal-sodium diet than those on low-sodium diet. As expected, our AN malnourished patients had lower fat mass than the healthy women. This was found in all of the numerous studies on body composition in AN patients.16e23 In the majority of them, body composition was assessed by anthropometry (skinfold
thickness), bioelectrical impedance analysis (BIA) or by dualenergy X-ray absorptiometry (DEXA). In one study,23 body composition was measured by the means of densitometry (underwater weighting). In all these studies, a dramatic fall in FFM was observed: in AN patients, the FFM was very significantly lower than in the control subjects.16e23 For example, Kerruish et al.19 found in 23 AN adolescent girls with a BMI of 15.4 1.3 kg/(m)2 (body weight: 40 kg), a FFM of 34.5 4.3 kg. Using underwater weighting in 238 AN patients ranging from 13 to 22 years of age and having a BMI between 11.2 and 17.4 kg/m2, Probst et al.23 found, using the Siri formula, a fat mass from 4.4% to 22.7% and a FFM from 77.3 to 95.6%. This represented a mean FFM of 35 kg, i.e. a theoretical minimal loss of 6 kg, compared with the mean FFM (80%) at the minimal BMI (18.5 kg/m2). Little is known concerning the body fluids in anorexia nervosa patients. Densitometry, anthropometry and DEXA do not permit to analyze body fluids (extra- and intracellular water). Unfortunately, we do not have the gold standard, in a clinical approach, for measurement of body water compartments in malnourished patients. In the study of Moreno et al.24 in 13 young AN patients (15 1.7 years; BMI: 14.5 1.4 kg/m2), the authors found different results than ours: they estimated 23.6 L for total body water, 13.2 L for intracellular water and 10.4 L for extracellular water. On the contrary, we found more ECW (15.4 kg) than ICW (13.5 kg). This may be due to the different devices used for body compartment measurement, to a group effect (only 13 patients were studied in the Moreno study) or to difference in AN subtypes. Indeed, in the Moreno study, only RAN patients were included,24 the mean BMI was 14.5 kg/m2 and the mean age 15.2 yr-old. Thus one can speculate that their patients had a BMI higher than the set-point under that the extracellular water inflation occurs. In the Mika study,25
754
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755
the 21 AN patients were 14.4 years old and the BMI 15.5 kg/m2. Their ECW was 9.7 L and their ICW 17.3 L. One more time, the difference with our own study could be related to the different device used (multifrequency analysis system) and to the BMI of these adolescent patients at the first measurement (extrapolated BMI for 23 years old: 17.4 kg/m2). Interestingly, the body weight gain of the AN patients in the present study was more precisely predicted by the energy balance when a low-sodium diet was prescribed than when the patients were on a normal-sodium diet. As an example, in the normalsodium diet group, the increase in body weight was between 1.1 and 2.4 kg per week, which was clearly higher than expected from the energy balance. In these patients, the “excess” in extracellular water reached 2.1 kg during the 3rde5th week of refeeding (see Fig. 2). In those of these patients having leg edema, the observed body weight gain was higher than 1.5 kg per week, which largely exceeded the expected body weight gain. This argues for the use of a low-sodium diet in these patients, in whom an accurate measurement of the “active weight” gain is needed. Indeed, in AN patients, cognitive behavioral therapy (CBT) integrates the energy needs for maintenance of body weight and the energy needs for body weight gain. In these patients, in whom the “fear of gaining too much body weight without eating so much” is very high, it should be necessary to have just the body weight gain that it was expected and explained to the patients. Importantly, in the present study, the prescription of this moderate low-sodium diet (4e5 g NaCl per day) was not associated with a lower energy intake by meals and snacks than in the normalsodium diet group: the energy intake by mouth was similar in the 2 groups. It must be noted that the excess body weight observed in our normal-sodium group could be higher if the patients would have only meals and snacks, without tube feeding, for gaining 1 kg/ week body weight. Thus, the fluid retention that we observed in our normal-sodium group could be an underestimation of that it would occur on a 2500e2800 kcal/d diet. The limits of this study should be remembered: there were not very severely malnourished AN patients, we did not include men, we used in all patients tube feeding associated with meals. More importantly, this was not a prospective randomized study. Indeed, we compared patients who were consecutively treated either with normal or low-sodium diet. Despite the fact that we did not change anything in the treatment of our patients during the period, we did not exclude that bias could interfere with our data: different dishes, different medications. Lastly, we have any means to know if the meals were consumed and if enteral mixtures were really administered. We checked the remaining mixtures and meals, but indeed it was not enough. Another limit was the use of a non validated method, BIA, for body composition, in these malnourished patients. We could just argue that we found a good correlation between BIA and skinfold thickness, as other authors have using BIA have found with the same confidence.17,26e28 These authors showed BIA correlated with skinfold thickness,17,26 to leptin level,27 an adipocyte hormone, or to nutritional status.28 In conclusion, the present study suggests that an increase in body extracellular water frequently occurs in AN patients under a BMI of 15e16 kg/m2. This moderated inflation in extracellular water disappears thereafter, during refeeding. It may be important to prescribe a low-sodium diet in order to avoid a body weight gain higher and faster than that expected by energy balance in these AN patients afraid of gaining too much weight. During the CBT sessions, the interest in this low-sodium diet, until the increase in urinary sodium output was reached, was explained to the patients. CBT takes also into account the composition of body weight gain, with an optimal repartition
between fat mass and fat-free mass. For all these reasons, it seems to us that, in malnourished or severely malnourished AN patients, it is useful to prescribe a low-sodium diet until the BMI will be higher than 15 kg/m2 in adult patients (which corresponds to a 14 kg/m2 for patients of 14e16 years old and 13 kg/m2 for 12e14 year-old adolescents, according to the Rolland-Cachera curves of BMI in infants). Conflict of interest statement No conflict of interest for any of the co-authors. Statement of authorship All authors agree to the publication. Acknowledgments The author acknowledges “association Autrement” (Dijon, France) and to the Ministère de la Santé, DGS, Paris, France References 1. Pinter O, Probst M, Vandereycken W, Pieters G, Goris M. The predictive value of body mass index for the weight evolution in anorexia nervosa. Eat Weight Disord 2004;9:232e5. 2. Coners H, Remschmidt H, Hebebrand. The relationship between premorbid body weight, weight loss, and weight at referral in adolescent patients with anorexia nervosa. Int J Eat Disord 1999;26:171e8. 3. Rigaud. Anorexie, boulimie, compulsions alimentaires. Les troubles du comportement alimentaire. Ed Marabout, Paris: p. 323. 4. Morgan HG, Hayward AE. Clinical assessment of anorexia nervosa. The MorganRussell outcome assessment schedule. Br J Psychiatry 1988;152:367e71. 5. Wild B, Friederich HC, Gross G, Teufel M, Herzog W, Giel KE, et al. The ANTOP study: focal psychodynamic psychotherapy, cognitive-behavioural therapy, and treatment-as-usual in outpatients with anorexia nervosaea randomized controlled trial. Trials 2009;10:23e33. 6. Godart N, Perdereau F, Galès O, Agman G, Deborde AS, Jeammet P. The weight contract during the hospitalization of anorexic patients. Arch Pediatr 2005;12:1544e50. 7. Allison S. Fluids, electrolytes and nutrition. Clin Med 2004;4:573e8. 8. Ehrlich S, Querfeld U, Pfeiffer E. Refeeding oedema: an important complication in the treatment of anorexia nervosa. Eur Child Adolesc Psychiatry 2006 Jun;15 (4):241e3. 9. Morgan HG, Barry R, Morgan MH. Myoedema in anorexia nervosa: a useful clinical sign. Eur Eat Disord Rev 2008;16:352e4. 10. De Caprio C, Zarrella L, Senatore I, Silvestri E, Contaldo F, Pasanisi F. [Clinical nutritional outcome in patients recovering in a psychiatric setting from severe protein-energy malnutrition of anorexia nervosa]. Ann Ital Med Int 2005 JuleSep;20(3):158e66. 11. Tey HL, Lim SC, Snodgrass AM. Refeeding oedema in anorexia nervosa. Singapore Med J 2005 Jun;46(6):308e10. 12. Rigaud D, Cohen B, Melchior JC, Moukaddem M, Reveillard V, Apfelbaum M. Refeeding improves muscle performances without normalization of muscle mass and oxygen consumption in anorexia nervosa. Am J Clin Nutr 1997;65:1845e51. 13. Van Wymelbeke V, Brondel L, Brun JM, Rigaud D. Factors linked to the increase in resting energy expenditure during refeeding in malnourished anorexia nervosa patients. Am J Clin Nutr 2004;80:1469e77. 14. Durning JVG, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1974;32:77e97. 15. Boulier A, Fricker J, Thomasset AL, Apfelbaum M. Fat-free mass estimation by the 2-electrode impedance method. Am J Clin Nutr 1990;52:581e5. 16. Marra M, Scalfi L, Caldara A, De Filippo E, Zurlo V, Contaldo F. Evaluation of body composition in anorexia nervosa. Minerva Gastroenterol Dietol 1997;43:143e8. 17. Piccoli A, Codognotto M, Di Pascoli L, Boffo G, Caregaro L. Body mass index and agreement between bioimpedance and anthropometry estimates of body compartments in anorexia nervosa. J Parenter Enteral Nutr 2005;29:148e56. 18. Scalfi L, Polito A, Bianchi L, Marra M, Caldara A, Nicolai E, et al. Body composition changes in patients with anorexia nervosa after complete weight recovery. Eur J Clin Nutr 2002;56:15e20. 19. Kerruish KP, O’Connor J, Humphries IR, Kohn MR, Clarke SD, Briody JN, et al. Body composition in adolescents with anorexia nervosa. Am J Clin Nutr 2002;75:31e7. 20. Bossu C, Galusca B, Normand S, Germain N, Collet P, Frere D, et al. Energy expenditure adjusted for body composition differentiates constitutional thinness from both normal subjects and anorexia nervosa. Am J Physiol Endocrinol Metab 2007 Jan;292(1):E132e7.
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755 21. Yamashita S, Kawai K, Yamanaka T, Inoo T, Yokoyama H, Morita C, et al. BMI, body composition, and the energy requirement for body weight gain in patients with anorexia nervosa. Int J Eat Disord; 2009 May 20. 22. Rigaud D, Brondel L, Poupard AT, Talonneau I, Brun JM. A randomized trial on the efficacy of a 2-month tube feeding regimen in anorexia nervosa: a 1-year follow-up study. Clin Nutr 2007 Aug;26(4):421e9. 23. Probst M, Goris M, Vandereycken W, Pieters G. Body composition in girls and young women with anorexia nervosa: comparison of different equations. Int J Eat Disord 2008;41:180e3. 24. Moreno MV, Djeddi DD, Jaffrin MY. Assessment of body composition in adolescent subjects with anorexia nervosa by impedance. Med Eng Phys 2008;30:783e91.
755
25. Mika C, Herpertz-Dahlmann B, Heer M, Holtkamp K. Improvement of nutritional status as assessed by multifrequency BIA during 15 weeks of refeeding in adolescent girls with anorexia nervosa. J Nutr 2004;134:3026e30. 26. Marra M, Caldara A, Montagnese C, De Filippo E, Pasanisi F, Contaldo F, et al. Bioelectrical impedance phase angle in constitutionally lean females, ballet dancers and patients with anorexia nervosa. Eur J Clin Nutr 2009;63:905e8. 27. Uzum AK, Yucel B, Omer B, Issever H, Ozbey NC. Leptin concentration indexed to fat mass is increased in untreated anorexia nervosa (AN) patients. Clin Endocrinol 2009;71:33e9. 28. Scalfi L, Marra M, Caldara A, Silvestri E, Contaldo F. Changes in bioimpedance analysis after stable refeeding of undernourished anorexic patients. Int J Obes Relat Metab Disord 1999;23:133e7.
Contents lists available at ScienceDirect
Clinical Nutrition journal homepage: http://www.elsevier.com/locate/clnu
Original Article
Body fluid retention and body weight change in anorexia nervosa patients during refeeding Daniel Rigaud a, *, Alain Boulier b, Isabelle Tallonneau a, Marie Claude Brindisi a, Raymond Rozen b a b
Service d’Endocrinologie et Nutrition, CHU Le Bocage, B.P. 77 908, 21079 Dijon, France Service d’explorations fonctionnellles, CHU Bichat, 75018 Paris, France
a r t i c l e i n f o
s u m m a r y
Article history: Received 16 December 2009 Accepted 28 May 2010
Background & aims: Body weight gain is an important goal in anorexia nervosa (AN) patients, but inflation in body fluids could artificially increase body weight during refeeding. Methods: 42 malnourished adult AN patients were refed using a normal-sodium diet, then 176 other malnourished adult AN patients received a refeeding low-sodium diet (BMI of the 218 patients: 13.4 1.9 kg/m2). Sodium balance, body composition by a 2-electrode impedance method (BIA, for assessment of total and extracellular water, fat-free mass, FFM), resting energy expenditure and energy intake were calculated. Results: In the patients on normal-sodium diet, body weight, and total and extracellular water gains were higher than those of the low-sodium diet patients (P < 0.01). Edema occurred more often in the former group (21% vs 6%; P < 0.05). In almost all patients, BMI reached a plateau around 15e16 kg/m2, then increased again. During this plateau, an increase in intracellular water and in “active FFM” was observed with BIA, together with a similar decrease in extracellular water. Conclusion: In AN patients, who are always afraid of gaining too much weight, in regard to their food intake, it will be useful to give a low-sodium diet until a 15e16 kg/m2 BMI. This should be integrated into the cognitive behavioral therapy. Ó 2010 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Keywords: Anorexia nervosa Body composition Eating disorder Malnutrition Tube feeding
1. Introduction In anorexia nervosa (AN), body weight is a key factor for medical decision. It has been described as one of the predictive factors for relapse.1e4 It has been included in different therapeutic goals and cognitive behavioral therapies.1,2,5 Reaching a normal BMI, for example a BMI at or above 18.5 kg/m2 in adult patients, seems a determinant for complete recovery. In AN patients, hospitalization is often decided because of body weight loss, as assessment of malnutrition. For some medical teams, body weight (BW) remains the primary goal of the contract signed by the hospitalized malnourished AN patients with the medical staff.6 In this “bodyweight contract”, some teams give to the patient’s authorization for familial visits or outdoor permissions according to body weight gain. In almost all the teams, the decision concerning the discharge from the hospital is determined by the BW.1,2,6 Lastly, when an AN patient is suspected to lie, because she doesn’t gain body weight despite high food and energy intake, it should be very important to * Corresponding author. Service d’Endocrinologie et Nutrition, CHU Le Bocage, 21000 Dijon, France. Tél.: þ33 3 80 29 34 52; fax: þ33 3 80 29 35 19. E-mail address: [email protected] (D. Rigaud).
know whether this lack of weight gain has to be interpreted as an absence of active mass gain, or as a gain in active mass associated with a loss in body water. Inversely, in the case of weight gain, the medical staff should know if this gain represents true active body mass, or only gain in fluids (and sodium) in extracellular compartments. Indeed, it is well known that, in several patients, malnutrition is associated with abnormalities of sodium-potassium pumps, inducing an increase in sodium and water retention.7 Legs or truncal edema are well described in malnourished non AN and AN patients,7,8 even if large edema are more frequent in Kwashiorkor than in AN malnourished patients. It is self evident that, in malnourished patients, body weight gain is partly composed of intracellular fluids in fat-free mass cells. But a gain in extracellular water could also be observed, in particular on free-sodium diet (10e12 g NaCl per day). Very few studies including only a small number of malnourished AN patients submitted to refeeding are published.8e11 Since we had observed that some patients, during refeeding, gained more body weight than expected from their energy balance, we decided to perform the present study to describe the changes in body weight, body masses and body fluids in 218 malnourished AN
0261-5614/$ e see front matter Ó 2010 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved. doi:10.1016/j.clnu.2010.05.007
750
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755
patients hospitalized for refeeding, either refed using a low-sodium or a normal-sodium diet. 2. Patients and methods 2.1. Patients In this study, the chart of all the consecutive malnourished adult AN women who were hospitalized for renutrition during the last 4 yrs in our Nutrition unit for at least 2 months was analyzed. The characteristics of these 218 malnourished AN patients were seen in Table 1. During the same period (4 yrs), 284 AN patients were hospitalized: 66 were not included in the analysis, because having a BMI >16 kg/(m)2 (n ¼ 8), being no compliant, refusing or stopping tube feeding, engaging frequent vomiting or purging episodes during stay, having too short stay (<2 months). In the first 42 patients, refeeding was performed using a combination of normal-sodium diet and nasogastric-tube feeding. In all, resting energy expenditure and body composition by bioimpedance were measured, as described before.12,13 Twenty four patients suffered from the restrictive subtype (RAN, 57%) and 18 others from the binge/purging subtype (BPAN) of the disease (DSM IV criteria). During their refeeding, there was no restriction in water or sodium intake from meals or snacks (normal-sodium diet group). The enteral mixture for tube feeding contained 0.6 g NaCl per liter (Sondalis isoÒ: 1000 mL ¼ 1000 kcal). Because, in our initial experience, several malnourished adult AN patients developed edemas or gained “too much” body weight, in regard of their low energy intake, we decided to use low-sodium diet in such patients. In the further 176 malnourished AN patients, was prescribed a low-sodium diet (4e5 g NaCl/day per 2000 kcal/ day). There were 92 RAN patients (56%) and 84 BPAN (44%). All had, as the previous 42 patients, meals and tube feeding for refeeding. The low-sodium diet was changed to a normal-sodium diet (10e12 g NaCl/day) as soon as the sodium urinary output has clearly increased. Among these 176 patients, 21 patients did not accept tube feeding anymore after a mean duration of 23 5 days of enteral nutrition. The criteria for non inclusion were diuretics or laxative misuse, corticotherapy, low compliance, or purging behavior. These AN patients were compared with 116 healthy women without eating disorder (Table 1). 2.2. Methods The following measurements were made every week in each patient:
2.2.1. Body composition 2.2.1.1. Anthropometry. Body height was measured at admission at the nearest lower 0.5 cm. Two times a week, body weight was measured at the nearest 100 g (reproducibility: 50 g), and BMI calculated. Skinfold thickness was determined at the 4 classical area (triceps, biceps, subscapular, suprailiac) using metallic Harpenden calliper and then body FM and FFM were calculated as follow: body density ¼ 1.1599[0.0717 (log S)], where S is the sum of the 4 skinfold.14 2.2.1.2. Bioelectrical impedance. It was used to measure body composition, as previously described.15 We obtained, using this two-frequencies analysis (impédancemètre Innerscan 2-fr BIA Tanita corp, France), the extracellular water (ECW) and body water (TW), then intracellular water (ICW) and fat-free mass (FFM), then active FFM, then fat mass (FM) from body weight. In this study, a double-frequency bioimpedance analyser15 with subcutaneous electrodes (using acupuncture needle with strict asepsis) was used, in order to obtain values for total body water (high frequency, 1 MHz) and extracellular water (low frequency, 50 kHz). One electrode was introduced under the skin of the right hand and the other under the skin of the left foot. Briefly, calculations were as following: Z2 ¼ R2 þ Xfc, Xfc¼(Z) sin q, R¼(Z) cos q; total body water (V): r H/S and V ¼ S H (R ¼ resistance, Xfc ¼ reactance, r ¼ resistivity, H ¼ height; FFM ¼ V/0.73; extracellular body water: V at 50 kHz; total body water ¼ V at 1 MHz, intracellular body water: totaleextracellular water; active fat-free mass ¼ FFMextracellular water; fat mass: body weight (measured just before BIA) minus fat-free mass. 2.2.2. Other measurements 2.2.2.1. Blood and urine analyses. Every week, total 48-h urinary output of sodium, potassium, urea and creatinine were measured; albumin, transthyrétine and hematocrit were obtained every two weeks. Energy intake was calculated from values of a food table (BilnutÒ, 1996), as previously described.12,13 We checked the remaining mixtures (tube feeding) and the remaining food after meals. In 39 patients on normal-sodium and 168 on low-sodium diet, resting energy expenditure (REE) was measured by indirect calorimetry.13 In the others, it was calculated from Harris-Benedict formulas. From these values, energy balance was calculated as follow: energy intake (from meals and tube feeding) minus energy needs. Energy needs during hospital stay were calculated as previously determined, i.e. resting energy expenditure 1.5 in patients having not excessive exercise and 1.7 in hyperactive patients. Theoretical mass retention (gain) was calculated on the
Table 1 Body composition at hospital’s admission in the 42 normal-sodium and the 176 low-sodium AN patients and in the 116 normal-weight healthy controls.
Age (years) Women (%) Body weight (kg) Body mass index (kg/(m)2) Total body water (kg) Total body water (%) Extracellular (EC, kg) Intracellular (IC, kg) EC/IC Fat-free mass (kg) Fat mass (kg) Active fat-free mass (kg) Fat-free mass (%) Fat mass (%)
Anorexia nervosa (n ¼ 42)
Anorexia nervosa (n ¼ 176)
Controls (n ¼ 116)
Expected body composition at 36 kga
22.1 4.2 97% 36.6 4.3 ** 13.8 1.7 ** 28.7 5.3 ** 78.6 3.0 ** 15.6 3.1 13.1 2.1 ** 1.19 0.16 ** 34.7 4.6 ** 1.9 1.7 ** 19.1 3.5 ** 94.8 5.1 ** 5.2 5.1 **
23.3 5.1 98% 36.0 3.8 ** 13.2 1.2 ** 28.8 4.8 ** 80.0 3.3 ** 15.3 2.8 13.5 2.5 ** 1.13 0.14 ** 34.1 4.6 ** 2.3 2.2 ** 18.8 3.1 ** 94.7 5.4 ** 5.3 5.1**
27.2 8.8 110 (95%) 57.6 9.3 22.3 2.6 35.6 4.7 61.8 3.4 16.1 2.1 19.5 2.7 0.82 0.05 44.6 7.1 13.0 4.3 31.2 4.9 77.4 4.1 22.6 4.1
e 36 4.4 13.4 1.9 23 2.4 63.9 3.6 10 1.4 13 1.8 0.77 0.06 29 2.4 7 1.7 19 2.2 80.6 4.5 19.4 3.1
**P < 0.005 vs controls; NS between AN patients. a As expected from body composition of the 116 control subjects at a body weight of 36 kg. Active fat-free mass: fat-free mass minus extracellular water.
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755
2.2.3. Statistical analyses Mean and standard deviation were given. From two-frequencies bioelectrical impedance (BIA), were measured total body water, extracellular water and fat-free mass. The active fat-free mass (CFFM) derived from fat-free mass, i.e. fat-free mass minus extracellular water (FFM ECW), as both obtained from BIA. From 48-h urinary sodium output and 48-h sodium intake (by meals and tube feeding) obtained during 2 consecutive days a week, we determined the total water body gain, using the following formula: 140 mmol Na þ gain, i.e. 9 g NaCl gain, corresponded to a gain of 1 L, i.e. one kilo of body weight. Creatinine urinary output was calculated as the mean of two successive 24-h urinary outputs. The results were analyzed according to BMI, with classes of BMI for each point of BMI (ex: 12e12.9, 13e13.9.). Student t test was used for comparisons between groups. Non parametric Wilcoxon tests were performed to compare two periods of time in our patients. 3. Results Tolerance of tube feeding was good: no severe side effect, no refeeding syndrome, no metabolic problem, 12 sinusitis (6%). All patients had normal hematocrit, transthyretin and albumin levels during all the study period (2 months). After 2 months, a decrease in albumin (from 42 to 38 g/L, P < 0.01) and in hematocrit level (from 41 to 38%, P < 0.01) was observed. The first 42 malnourished AN patients were refed using a free normal-sodium diet (3 meals and 1 or 2 snacks) and tube feeding, in order to complement energy input. The increase in body weight was 1.42 0.23 kg after one week of refeeding, despite an energy intake around their energy needs (1638 234 kcal/day). In 9 out of the 42 patients (21%), all having a BMI<15 kg/(m)2, indisputable legs edema were observed during renutrition. This was associated with a mean body weight gain of 1.34 0.31 kg per week during the first 3 weeks of refeeding. In two of them, having the BPAN subtype, the edema were large. In 8 of these 9 patients, the edema disappeared when their BMI reached a value between 15 and 17 kg/(m)2. In the last one (BPAN patient), edema persisted until the discharge from the hospital. The further 176 malnourished AN patients were refed using a free (as they could eat) but low-sodium diet including 3 meals and 1 or 2 snacks. Renutrition also included tube feeding. Their body weight increase was 0.73 0.18 kg after one week of refeeding, with an energy intake around their energy needs (1583 207 kcal/ day). The low-sodium diet was given until the urinary sodium output clearly increased. This increase in urinary output often occurred after 4e6 weeks of refeeding. In 11 patients (6%; BPAN: n ¼ 8), all having a BMI<15 kg/(m)2, light edema of the legs were observed during renutrition. Edema were less frequent than in the normal-sodium diet group (chi square; P < 0.05). The patients on low-sodium diet gained 9.3 0.9 kg of body weight during the 2-month refeeding. BMI increased from 12.6 1.2 to 17.6 1.2 kg/(m)2. After one month refeeding, energy input reached 2448 162 kcal/day (including 886 115 kcal by tube feeding) in the normal-sodium diet group and 2512 171 kcal/day (including 924 121 kcal by tube feeding) in the low-sodium diet group. After 2 months, energy input reached
2684 186 kcal/day (812 109 kcal by tube feeding) in the normalsodium diet group and 2647 192 kcal/day (862 92 kcal by tube feeding) in the low-sodium diet group. There was no difference in these energy input between the low- and the normal-sodium group (ANOVA, F ¼ 0.87). Until a mean BMI of 15e16 kg/(m)2, the mean BMI of the 2 groups (low- and normal-sodium diets) increased linearly. Around a BMI value of 15.2 0.4 kg/(m)2, the body weight remained stable during 7e10 days. This plateau occurred at the threshold of BMI 15e16 kg/(m)2 in 192 patients (88%). This stability of body weight contrasted with the absence of any decrease in energy intake, which was not significantly different from those of the previous and the next weeks. Before it reached this 15e16 kg/(m)2 value, BMI of the patients in the normal-sodium diet increased more than BMI of the low-sodium diet patients (P < 0.05). This occurred despite similar resting energy expenditure and similar energy input by meals and tube feeding in the 2 groups (see above). 3.1. Sodium urinary output The course of sodium urinary output was seen in Fig. 1. In the 34 normal-sodium diet patients for whom we had sodium daily balance, the sodium balance was positive until the BMI reached a value of 15e16 kg/(m)2, then it tended to be negative. In the 71 patients on low-sodium diet for whom we had sodium daily balance, the sodium balance was less positive (difference between groups, P < 0.01). In the two groups, mean sodium balance became moderately negative, when BMI reached a value of 15e16 kg/(m)2. 3.2. Body composition with the two-frequency bioelectrical impedance analysis (2-fr BIA) Table 1 gives the results of BIA in the AN patients, compared with the healthy subjects. The AN patients had higher total body water and higher extracellular water (P < 0.01). Total body water represented 78.4 3.2% of body weight in the AN patients and 63.9 3.6% in the controls (P < 0.01). The extra-/intracellular water ratio was higher in AN patients than in controls. Compared with values extrapolated from control values at a 36 kg-body weight, the AN patients had lower fat mass (P < 0.01), and higher fat-free mass than these shadow controls (P < 0.01), but similar active FFM (NS).
30
25
20
15 mmol/24 hr
basis of 7.5 kcal/g body weight storage. This allowed us to expect the body weight gain by week, and to conclude whether or not the body weight gain was in accord with the energy-excess. The protocol of this study was approved by the ethical committee of Paris 7 university. Indirect calorimetry (for REE) and bioelectrical impedance (for body composition) were performed in all patients admitted in our unit in this period, even those who did not participate in the present study. Each patient gave informed consent to participate.
751
10
*
*
*
*
11-12
12-13
13-14
14-15
5
0
-5
-10 15-16
16-17
17-18 2
Normal sodium diet
low sodium diet
BMI (kg/m )
Fig. 1. Evolution of daily sodium balance in 34 patients refed by normal-sodium diet and in 71 patients on low-sodium diet. Normal-sodium diet: sodium intake 10.3 2.3 g NaCl/day; low-sodium diet: 4.3 1.2 g NaCl/day (including tube feeding: 600 mg/L [1000 kcal]). The fact that sodium balance was going to be negative from P < 0.05. a BMI value of 15e16 kg/m2 must be noted.
752
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755
Table 2 Changes in body composition during refeeding in the 176 anorexia nervosa patients refed with a low-sodium diet, according to the type of the disease.
Body weight (kg) BMI (kg/(m)2) Total body water (kg) Extracellular W (kg) Intracellular W (kg) Extra-/intra water ratio Fat-free mass (kg) Fat mass (kg) Active FFM (kg)
RAN (n ¼ 92) admission
RAN (n ¼ 90) 2 months
BPAN (84) admission
BPAN (n ¼ 81) 2 months
35.2 4.1 13.1 1.2 26.9 2.1 14.5 1.3 12.4 1.0 1.17 0.11 33.4 1.6 1.8 0.5 18.9 1.1
44.4 2.1 16.8 1.4 29.7 1.9 13.3 1.1 16.5 1.5 0.80 0.08 $ 39.1 1.7 5.3 0.6 25.8 1.5
38.4 3.7 * 14.3 0.7 * 24.5 1.8 12.3 1.2 * 12.2 1.5 1.01 0.12 * 35.1 2.8 3.7 1.5 * 22.8 1.8 *
48.4 2.5 17.9 1.8 28.1 2.0 12.6 1.2 15.5 1.5 0.81 0.10 42.5 2.0 7.0 0.8 29.9 1.3
RAN: restrictive subtype, BPAN: binge-eating/purging subtype. W: water. Results were obtained using two-frequencies bioelectrical impedance analysis at admission and after a 2-month refeeding (mean of 57 6 days). A low-sodium diet (<4 g NaCl/day) was given. As shown, despite a body weight gain of 8e10 kg, extracellular water did not increase. *P < 0.02 versus RAN (patients suffering from self-induced vomiting). In the 2 groups, the values obtained at 2 months were significantly higher than those obtained before the start of refeeding (P < 0.001), except for extracellular water and EW/IW ratio.
This suggested that the water content of FFM was too high in AN patients. Table 2 gives the results of BIA in the AN patients, according to the type of the disease. The patients having the binge-eating/ purging subtype (BPAN) had higher body weight (and BMI), higher FFM and higher FM than the RAN patients (P < 0.01). For this higher body weight, the BPAN patients had lower total body water and lower extracellular water (non significant) than the RAN patients. As shown in Table 2, the extra-/intracellular water ratio, which was elevated before refeeding, decreased with refeeding, while the patients had reached a BMI higher than 15 kg/(m)2. After 2-month refeeding, only two variables were not increased: extracellular water and extra-/intracellular ratio. Body composition changes differed between the normalsodium and the low-sodium diets (Table 3). On normal-sodium diet, the patients gained more body weight, more FFM than the patients on low-sodium diet (P < 0.05). As shown in Fig. 2, the increase in body weight during refeeding (P < 0.001) was in part due to an increase in extracellular fluids until the fourth week (P < 0.02). The curve of active FFM and intracellular water were linear, although that of extracellular water was omega shaped (maximal values between week 3 and week 5 in the two groups). The active FFM (FFM minus the extracellular water) and the intracellular water tended to increase at a constant rate, while extracellular water did not. Similarly, between the 4th and the 6th weeks, while energy input did not decrease, body weight and FFM did not increase anymore, for one to two weeks. This corresponded to a BMI value between 15 and 16 kg/(m)2. During the same period, fat mass, ICW and active FFM increased
linearly in the 2 diet groups (Fig. 2). By contrast, extracellular water showed different evolution in the two diet groups: ECW did not change in the low-sodium group, while it increased then decreased in the normal-sodium diet group. 3.3. Relationship between skinfold thickness and 2-fr-bioelectrical impedance analysis As determined from the four skinfold thickness, FM increased more using skinfold thickness measurement than using BIA. The gain of FM after one month refeeding was þ1.8 0.2 kg using skinfold thickness versus þ1.2 0.3 kg using BIA in the normalsodium group and þ1.8 0.2 kg (skinfold thickness) vs þ1.2 0.2 kg using BIA in the low-sodium diet group (P < 0.05 for each). There was a good correlation for FM between the values obtained from skinfold thickness and BIA (r ¼ 0.684, P < 0.001). But BIA systematically underestimated FM. For example, 18 out of our 218 patients (8%) had a BMI lower than 11 kg/(m)2: in 14 of them (77%), fat mass was negative in BIA (1.1 0.4 kg), because BIAcalculated FFM was higher than body weight. In these 14 patients, FM was þ0.9 0.2 kg using skinfold thickness. 3.4. Relationship between energy expenditure and body weight gain Considering the mean energy balance (energy intake minus daily energy expenditure) during the first month of refeeding, the expected body weight gain of the normal-sodium diet patients was
Table 3 Changes in body composition during refeeding in the normal-sodium and low-sodium diet groups, using skinfold thickness or BIA measurement.
BMI (normal-sodium diet) BMI (low-sodium diet) Weight (normal-sodium diet) Weight (low-sodium diet) FFM (NSD) skinfold FFM (low SD) skinfold FMM (NSD) BIA FMM (low SD) BIA Fat mass (NSD) skinfold Fat mass (low SD) skinfold Fat mass (NSD) BIA Fat mass (low SD) BIA
Week 0
W2
W4
W6
W8
13.3 1.9 13.7 1.6 36.2 4.3 36.5 2.3 33.3 2.8 33.4 2.0 34.1 3.7 33.9 3.3 2.9 1.3 3.1 1.1 2.1 1.5 2.6 1.4
14.5 1.6 14.7 1.3 39.4 3.4 38.9 2.1 35.9 2.0 35.1 1.7 35.6 3.5 36.0 2.9 3.5 1.0 3.8 1.0 2.8 1.2 2.9 1.3
15.9 1.3 15.0 1.6 43.4 3.1 39.9 2.0 39.3 2.4 35.6 1.8 * 37.0 3.6 36.8 2.9 4.1 1.1 4.3 1.1 3.7 1.2 3.1 1.2
15.6 1.5 15.9 1.7 42.5 3.3 41.9 2.1 * 38.0 2.2 36.9 1.7 * 38.9 3.7 38.4 2.8 * 4.6 1.0 4.9 0.9 3.9 1.3 3.6 1.1
16.2 1.6 16.4 1.5 44.2 3.1 43.6 2.2 * 38.9 2.5 38.1 1.8 * 38.3 3.5 39.4 2.7 * 5.3 0.9 5.6 0.8 4.0 1.4 4.2 1.2
Measurements were performed each week in a mean of 36 patients on normal-sodium diet and in a mean of 164 patients on low-sodium diet. BIA: 2-fr-bioelectrical impedance analysis. Skinfold: from triceps, biceps, subscapular, suprailiac measurements. For skinfold thickness, FFM was calculated as body weight minus fat mass. For both skinfold and BIA calculation, FFM included extracellular water. * significantly different from normal-sodium diet (NSD): P < 0.05.
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755
753
28
26
24
22
kg
active FFM 20
normal-sodium diet
*
*
18
extracellular water
*
16
low-sodium diet 14
intracellular water 12 W0 active FMM NSD
W1
W2
W3
active FMM low SD
W4 ECW NSD
W5 ECW low SD
W6
W7
W8 weeks
ICW NSD
ICW low SD
Fig. 2. Evolution of active FFM, extracellular and intracellular water in the patients on normal-sodium or low-sodium diet Squares: normal-sodium water group; circles: lowsodium water group. Black: FFM, grey: extracellular water, white: intracellular water. The patients were refed both by meals and tube feeding. Active fat-free mass (FFM minus extracellular water) and intracellular water increased at the same rhythm in the 2 groups. By contrast, extracellular water increased more in the normal-sodium diet group. Notes P < 0.05 at W3, W4 and W5 between normalthat the curves of intra- and extracellular water crossed each other at W5-W6 in both groups (mean BMI: 15.34 0.28 kg/(m)2). sodium and low-sodium diet groups.
3.9 0.3 kg, while it was measured at 6.7 0.5 kg (P < 0.01 vs expected weight gain). In the low-sodium diet group, the expected weight gain was 4.0 0.4 kg and the true gain 3.8 0.3 kg (expected vs true: not significant). In the two groups, after that the BMI had reached the value of 15 kg/(m)2, the weight gain was similar as that expected. 4. Discussion In these malnourished AN patients, a BMI and FFM plateau was observed during refeeding, despite unchanged energy intake or expenditure. This plateau occurred although the patients did not reduce their energy intake. It was observed between the 4th and the 6th week of hospitalization. This plateau occurred very frequently at a 15e16 kg/(m)2 BMI. At this time, we observed a tendency for a decrease in total FFM and extracellular water, while active FFM (and intracellular water) and fat mass continued to increase. During the same period, the sodium balance, which was clearly positive during the previous weeks (in particular in the normal-sodium diet group) turned to be negative during 2e3 weeks. This was related to an increase in sodium urinary output, and not to a decrease in sodium intake. The increase in sodium loss associated with the decrease in extracellular water suggested that, at this 15e16 BMI value, the sodium and water retention related to malnutrition tended to disappear. These changes were more accurate in patients submitted to normal-sodium diet than those on low-sodium diet. As expected, our AN malnourished patients had lower fat mass than the healthy women. This was found in all of the numerous studies on body composition in AN patients.16e23 In the majority of them, body composition was assessed by anthropometry (skinfold
thickness), bioelectrical impedance analysis (BIA) or by dualenergy X-ray absorptiometry (DEXA). In one study,23 body composition was measured by the means of densitometry (underwater weighting). In all these studies, a dramatic fall in FFM was observed: in AN patients, the FFM was very significantly lower than in the control subjects.16e23 For example, Kerruish et al.19 found in 23 AN adolescent girls with a BMI of 15.4 1.3 kg/(m)2 (body weight: 40 kg), a FFM of 34.5 4.3 kg. Using underwater weighting in 238 AN patients ranging from 13 to 22 years of age and having a BMI between 11.2 and 17.4 kg/m2, Probst et al.23 found, using the Siri formula, a fat mass from 4.4% to 22.7% and a FFM from 77.3 to 95.6%. This represented a mean FFM of 35 kg, i.e. a theoretical minimal loss of 6 kg, compared with the mean FFM (80%) at the minimal BMI (18.5 kg/m2). Little is known concerning the body fluids in anorexia nervosa patients. Densitometry, anthropometry and DEXA do not permit to analyze body fluids (extra- and intracellular water). Unfortunately, we do not have the gold standard, in a clinical approach, for measurement of body water compartments in malnourished patients. In the study of Moreno et al.24 in 13 young AN patients (15 1.7 years; BMI: 14.5 1.4 kg/m2), the authors found different results than ours: they estimated 23.6 L for total body water, 13.2 L for intracellular water and 10.4 L for extracellular water. On the contrary, we found more ECW (15.4 kg) than ICW (13.5 kg). This may be due to the different devices used for body compartment measurement, to a group effect (only 13 patients were studied in the Moreno study) or to difference in AN subtypes. Indeed, in the Moreno study, only RAN patients were included,24 the mean BMI was 14.5 kg/m2 and the mean age 15.2 yr-old. Thus one can speculate that their patients had a BMI higher than the set-point under that the extracellular water inflation occurs. In the Mika study,25
754
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755
the 21 AN patients were 14.4 years old and the BMI 15.5 kg/m2. Their ECW was 9.7 L and their ICW 17.3 L. One more time, the difference with our own study could be related to the different device used (multifrequency analysis system) and to the BMI of these adolescent patients at the first measurement (extrapolated BMI for 23 years old: 17.4 kg/m2). Interestingly, the body weight gain of the AN patients in the present study was more precisely predicted by the energy balance when a low-sodium diet was prescribed than when the patients were on a normal-sodium diet. As an example, in the normalsodium diet group, the increase in body weight was between 1.1 and 2.4 kg per week, which was clearly higher than expected from the energy balance. In these patients, the “excess” in extracellular water reached 2.1 kg during the 3rde5th week of refeeding (see Fig. 2). In those of these patients having leg edema, the observed body weight gain was higher than 1.5 kg per week, which largely exceeded the expected body weight gain. This argues for the use of a low-sodium diet in these patients, in whom an accurate measurement of the “active weight” gain is needed. Indeed, in AN patients, cognitive behavioral therapy (CBT) integrates the energy needs for maintenance of body weight and the energy needs for body weight gain. In these patients, in whom the “fear of gaining too much body weight without eating so much” is very high, it should be necessary to have just the body weight gain that it was expected and explained to the patients. Importantly, in the present study, the prescription of this moderate low-sodium diet (4e5 g NaCl per day) was not associated with a lower energy intake by meals and snacks than in the normalsodium diet group: the energy intake by mouth was similar in the 2 groups. It must be noted that the excess body weight observed in our normal-sodium group could be higher if the patients would have only meals and snacks, without tube feeding, for gaining 1 kg/ week body weight. Thus, the fluid retention that we observed in our normal-sodium group could be an underestimation of that it would occur on a 2500e2800 kcal/d diet. The limits of this study should be remembered: there were not very severely malnourished AN patients, we did not include men, we used in all patients tube feeding associated with meals. More importantly, this was not a prospective randomized study. Indeed, we compared patients who were consecutively treated either with normal or low-sodium diet. Despite the fact that we did not change anything in the treatment of our patients during the period, we did not exclude that bias could interfere with our data: different dishes, different medications. Lastly, we have any means to know if the meals were consumed and if enteral mixtures were really administered. We checked the remaining mixtures and meals, but indeed it was not enough. Another limit was the use of a non validated method, BIA, for body composition, in these malnourished patients. We could just argue that we found a good correlation between BIA and skinfold thickness, as other authors have using BIA have found with the same confidence.17,26e28 These authors showed BIA correlated with skinfold thickness,17,26 to leptin level,27 an adipocyte hormone, or to nutritional status.28 In conclusion, the present study suggests that an increase in body extracellular water frequently occurs in AN patients under a BMI of 15e16 kg/m2. This moderated inflation in extracellular water disappears thereafter, during refeeding. It may be important to prescribe a low-sodium diet in order to avoid a body weight gain higher and faster than that expected by energy balance in these AN patients afraid of gaining too much weight. During the CBT sessions, the interest in this low-sodium diet, until the increase in urinary sodium output was reached, was explained to the patients. CBT takes also into account the composition of body weight gain, with an optimal repartition
between fat mass and fat-free mass. For all these reasons, it seems to us that, in malnourished or severely malnourished AN patients, it is useful to prescribe a low-sodium diet until the BMI will be higher than 15 kg/m2 in adult patients (which corresponds to a 14 kg/m2 for patients of 14e16 years old and 13 kg/m2 for 12e14 year-old adolescents, according to the Rolland-Cachera curves of BMI in infants). Conflict of interest statement No conflict of interest for any of the co-authors. Statement of authorship All authors agree to the publication. Acknowledgments The author acknowledges “association Autrement” (Dijon, France) and to the Ministère de la Santé, DGS, Paris, France References 1. Pinter O, Probst M, Vandereycken W, Pieters G, Goris M. The predictive value of body mass index for the weight evolution in anorexia nervosa. Eat Weight Disord 2004;9:232e5. 2. Coners H, Remschmidt H, Hebebrand. The relationship between premorbid body weight, weight loss, and weight at referral in adolescent patients with anorexia nervosa. Int J Eat Disord 1999;26:171e8. 3. Rigaud. Anorexie, boulimie, compulsions alimentaires. Les troubles du comportement alimentaire. Ed Marabout, Paris: p. 323. 4. Morgan HG, Hayward AE. Clinical assessment of anorexia nervosa. The MorganRussell outcome assessment schedule. Br J Psychiatry 1988;152:367e71. 5. Wild B, Friederich HC, Gross G, Teufel M, Herzog W, Giel KE, et al. The ANTOP study: focal psychodynamic psychotherapy, cognitive-behavioural therapy, and treatment-as-usual in outpatients with anorexia nervosaea randomized controlled trial. Trials 2009;10:23e33. 6. Godart N, Perdereau F, Galès O, Agman G, Deborde AS, Jeammet P. The weight contract during the hospitalization of anorexic patients. Arch Pediatr 2005;12:1544e50. 7. Allison S. Fluids, electrolytes and nutrition. Clin Med 2004;4:573e8. 8. Ehrlich S, Querfeld U, Pfeiffer E. Refeeding oedema: an important complication in the treatment of anorexia nervosa. Eur Child Adolesc Psychiatry 2006 Jun;15 (4):241e3. 9. Morgan HG, Barry R, Morgan MH. Myoedema in anorexia nervosa: a useful clinical sign. Eur Eat Disord Rev 2008;16:352e4. 10. De Caprio C, Zarrella L, Senatore I, Silvestri E, Contaldo F, Pasanisi F. [Clinical nutritional outcome in patients recovering in a psychiatric setting from severe protein-energy malnutrition of anorexia nervosa]. Ann Ital Med Int 2005 JuleSep;20(3):158e66. 11. Tey HL, Lim SC, Snodgrass AM. Refeeding oedema in anorexia nervosa. Singapore Med J 2005 Jun;46(6):308e10. 12. Rigaud D, Cohen B, Melchior JC, Moukaddem M, Reveillard V, Apfelbaum M. Refeeding improves muscle performances without normalization of muscle mass and oxygen consumption in anorexia nervosa. Am J Clin Nutr 1997;65:1845e51. 13. Van Wymelbeke V, Brondel L, Brun JM, Rigaud D. Factors linked to the increase in resting energy expenditure during refeeding in malnourished anorexia nervosa patients. Am J Clin Nutr 2004;80:1469e77. 14. Durning JVG, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1974;32:77e97. 15. Boulier A, Fricker J, Thomasset AL, Apfelbaum M. Fat-free mass estimation by the 2-electrode impedance method. Am J Clin Nutr 1990;52:581e5. 16. Marra M, Scalfi L, Caldara A, De Filippo E, Zurlo V, Contaldo F. Evaluation of body composition in anorexia nervosa. Minerva Gastroenterol Dietol 1997;43:143e8. 17. Piccoli A, Codognotto M, Di Pascoli L, Boffo G, Caregaro L. Body mass index and agreement between bioimpedance and anthropometry estimates of body compartments in anorexia nervosa. J Parenter Enteral Nutr 2005;29:148e56. 18. Scalfi L, Polito A, Bianchi L, Marra M, Caldara A, Nicolai E, et al. Body composition changes in patients with anorexia nervosa after complete weight recovery. Eur J Clin Nutr 2002;56:15e20. 19. Kerruish KP, O’Connor J, Humphries IR, Kohn MR, Clarke SD, Briody JN, et al. Body composition in adolescents with anorexia nervosa. Am J Clin Nutr 2002;75:31e7. 20. Bossu C, Galusca B, Normand S, Germain N, Collet P, Frere D, et al. Energy expenditure adjusted for body composition differentiates constitutional thinness from both normal subjects and anorexia nervosa. Am J Physiol Endocrinol Metab 2007 Jan;292(1):E132e7.
D. Rigaud et al. / Clinical Nutrition 29 (2010) 749e755 21. Yamashita S, Kawai K, Yamanaka T, Inoo T, Yokoyama H, Morita C, et al. BMI, body composition, and the energy requirement for body weight gain in patients with anorexia nervosa. Int J Eat Disord; 2009 May 20. 22. Rigaud D, Brondel L, Poupard AT, Talonneau I, Brun JM. A randomized trial on the efficacy of a 2-month tube feeding regimen in anorexia nervosa: a 1-year follow-up study. Clin Nutr 2007 Aug;26(4):421e9. 23. Probst M, Goris M, Vandereycken W, Pieters G. Body composition in girls and young women with anorexia nervosa: comparison of different equations. Int J Eat Disord 2008;41:180e3. 24. Moreno MV, Djeddi DD, Jaffrin MY. Assessment of body composition in adolescent subjects with anorexia nervosa by impedance. Med Eng Phys 2008;30:783e91.
755
25. Mika C, Herpertz-Dahlmann B, Heer M, Holtkamp K. Improvement of nutritional status as assessed by multifrequency BIA during 15 weeks of refeeding in adolescent girls with anorexia nervosa. J Nutr 2004;134:3026e30. 26. Marra M, Caldara A, Montagnese C, De Filippo E, Pasanisi F, Contaldo F, et al. Bioelectrical impedance phase angle in constitutionally lean females, ballet dancers and patients with anorexia nervosa. Eur J Clin Nutr 2009;63:905e8. 27. Uzum AK, Yucel B, Omer B, Issever H, Ozbey NC. Leptin concentration indexed to fat mass is increased in untreated anorexia nervosa (AN) patients. Clin Endocrinol 2009;71:33e9. 28. Scalfi L, Marra M, Caldara A, Silvestri E, Contaldo F. Changes in bioimpedance analysis after stable refeeding of undernourished anorexic patients. Int J Obes Relat Metab Disord 1999;23:133e7.