Body composition, eating disorder psychopathology, and psychological distress in anorexia nervosa: a longitudinal study1–3 Marwan El Ghoch, Chiara Milanese, Simona Calugi, Massimo Pellegrini, Nino Carlo Battistini, and Riccardo Dalle Grave
INTRODUCTION
Anorexia nervosa (AN)4 is characterized by a restriction of energy intake relative to that required, which causes sufferers to become significantly underweight (1). Both the symptoms associated with low body weight and dietary restriction (starvation symptoms) and the psychopathology behind the eating disorder (2) can be addressed through nutritional rehabilitation (3), weight restoration, and cognitive behavioral therapy (4). These strategies are also necessary to restore an adequate quantity of body mass, both lean and fat, to enable physiologic processes (ie, a menstrual cycle and resting energy expenditure) to resume (5).
As to be expected from their undereating and underweight conditions, a marked reduction in both fat and lean masses in individuals with AN has been well documented (6, 7). Several studies have examined changes in body composition and body fat distribution in adults with AN. Of these studies, the majority of trials showed a preferential trunk fat accumulation associated with an acute weight regain immediately after treatment (8–11), although this was not a universal finding (12). Longer term, one study showed no differences in visceral adipose tissue distribution between former patients with AN who had maintained normal weight for 1 y and BMI-matched control women with no history of AN (13). However, a lower percentage of body fat restoration after weight restoration does seem to be a risk factor for relapse in patients with AN (14, 15). Nevertheless, the research conducted to date on this issue has suffered from various methodologic limitations, such as small sample groups (8, 9), incomplete weight restoration (8, 9), and small (8, 10, 11, 13) and unsuitable control groups (normalweight but not lean participants) (12). Thus far, research has also failed to address the relation between fat distribution, eating disorder psychopathology, and the psychological distress associated with weight restoration in AN (16). Indeed, although some authors have hypothesized that patients with greater trunk fat deposition after weight restoration could experience considerable stress brought on by changes in their body shapes, which make them more prone to a relapse (10), data to support this theory are lacking. A reported association between depressive mood and visceral adipose tissue in overweight women (17), possibly involving the leptin-mediated dysregulation of the
1 From the Department of Eating and Weight Disorders, Villa Garda Hospital, Garda, Italy (MEG, SC, and RDG); the Department of Neurological and Movement Sciences, University of Verona, Verona, Italy (CM); and the Department of Diagnostic, Clinical and Public Health, University of Modena and Reggio Emilia, Modena and Reggio Emilia, Italy (MP and NCB). 2 No support was received for this study. 3 Address correspondence to M El Ghoch, Department of Eating and Weight Disorders, Villa Garda Hospital, Via Montebaldo, 89, I-37016 Garda, Verona, Italy. E-mail: [email protected]. 4 Abbreviations used: AN, anorexia nervosa; BMC, bone mineral content; BSI, Brief Symptom Inventory; BSI-GSI, Global Severity Index of the Brief Symptom Inventory; CBT-E, enhanced form of cognitive behavioral therapy; DXA, dual-energy X-ray absorptiometry; EDE, Eating Disorder Examination interview; GSI, Global Severity Index. Received October 25, 2013. Accepted for publication January 21, 2014. First published online February 5, 2014; doi: 10.3945/ajcn.113.078816.
Am J Clin Nutr 2014;99:771–8. Printed in USA. Ó 2014 American Society for Nutrition
771
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
ABSTRACT Background: Although the effect of immediate weight restoration on body composition and body fat distribution has previously been studied in anorexia nervosa (AN), its influence in women with AN on eating disorder psychopathology and psychological distress has not previously been investigated to our knowledge. Objectives: We assessed body composition and fat mass distribution before and after body weight restoration and investigated any relation between changes in body fat patterns of patients with AN treated in a specialist inpatient unit and their eating disorder and psychological distress features. Design: Body composition was measured by using dual-energy X-ray absorptiometry in 50 female, adult patients with AN before and after complete weight restoration [body mass index (BMI; in kg/m2) $18.5] and 100 healthy control subjects matched by age and posttreatment BMI of study group participants. Eating disorder psychopathology and psychological distress were assessed in the AN group before and after weight restoration by using the Eating Disorder Examination interview and the Global Severity Index of the Brief Symptom Inventory (BSI-GSI), respectively. Results: After the achievement of complete weight restoration, patients with AN had higher trunk (P , 0.001), android (P , 0.001), and gynoid (P , 0.001) fat masses and lower arm (P , 0.001) and leg (P = 0.001) fat masses with respect to control subjects. No relation was shown between body-composition variables and eating disorder psychopathology in the AN group, and the only significant predictor of change in BSI-GSI was the baseline BSI-GSI score. Conclusion: The normalization of body weight in patients with AN is associated with a preferential distribution of body fat in central regions, which does not, however, seem to influence either eating disorder psychopathology or psychological distress scores. This trial was registered at www.controlled-trials.com as ISRCTN92626057. Am J Clin Nutr 2014;99:771–8.
772
EL GHOCH ET AL
SUBJECTS AND METHODS
Participants Fifty white, female patients with AN and 100 normal-weight, white, female subjects were enrolled in the study. All 50 patients were voluntarily and consecutively admitted to the eating disorder inpatient unit of Villa Garda Hospital during the years 2011–2013. Inclusion criteria were as follows: 1) aged 18–50 y, 2) BMI (in kg/m2) #17.5, 3) Eating Disorder Examination interview (EDE; 12.0D) diagnosis of AN except the amenorrhea criterion (21, 22), 4) either failure of less-intensive outpatient treatment or an eating disorder of clinical severity that was not manageable in an outpatient setting, and 5) BMI $18.5 at the end of the 20-wk treatment. Patients with active substance abuse, schizophrenia, and other psychotic disorders were excluded. Demographic and baseline clinical characteristics of patients with AN are presented in Table 1. Patients with AN were carefully matched with 100 control individuals of the same age (62 y) with a BMI equivalent to patients’ BMI at the end of their treatment. Control subjects were recruited randomly from university listservs and advertisements and were all healthy and weight-stable, with regular menstruation, no history of eating disorders or other significant psychiatric or medical conditions, and no current psychotropic medication intake. Inclusion criteria for both patients and controls were evaluated during an eligibility interview conducted by a specialist in the field (RDG). The study design was reviewed and approved by the Institutional Review Board of Villa Garda Hospital, Verona, and all participants gave written informed consent for the use of their anonymous personal data.
TABLE 1 Demographic and baseline clinical characteristics in patients with AN and the control group1
Age (y) Highest educational qualification [n (%)] Elementary school Middle school certificate High school diploma Bachelor’s degree Missing data Presence of amenorrhea [n (%)] Duration of illness (y)3 Amount of exercise (min)3,4
Pretreatment AN (n = 50)
Control group (n = 100)
26.5 6 7.62
26.5 6 7.6
0 13 (26.0) 26 (52.0) 9 (18.0) 2 (4.0) 42 (84.0) 8.7 6 6.7 1533.9 6 2732.3
0 0 50 (50.0) 50 (50.0) 0 0 (0.0) — —
1
AN, anorexia nervosa. Mean 6 SD (all such values). 3 Data were not available for the control group. 4 In the past 28 d (measured by using the Eating Disorder Examination interview, 12th ed). 2
Inpatient treatment protocol The inpatient treatment has been described in detail elsewhere (23). In brief, the treatment was an adapted version of the enhanced form of cognitive behavioral therapy (CBT-E) for eating disorders (4) that lasts 20 wk and comprises 13 wk of inpatient therapy followed by 7 wk of partial hospitalization. The program retained all main strategies and procedures of the CBT-E, which are delivered in both individual and group sessions. However, the program differed from the outpatient CBT-E in that patients received dietitian-assisted eating in the early weeks of treatment until their BMI reached $18.5. Over the first 3 wk of treatment, the dietary intake of patients was increased incrementally from 1500 to 2500 kcal. In subsequent weeks of the program, the dietary energy content was adjusted to maintain a weight gain rate from 1 to 1.5 kg/wk until the BMI of patients reached $19. Once this target was achieved, the calorie content of meals was continually adjusted so that body weight of patients remained within a 2-kg range of this figure. The diet included all main food groups and conformed to Italian National Guidelines for Healthy Eating (24). In addition, patients with BMI $15 and stable medical conditions participated in physiotherapist-led 60-min physical exercise sessions 2 times/wk as part of their treatment. These sessions included calisthenic exercises to help restore muscle strength, flexibility, and posture as well as aerobic exercises to improve cardiovascular fitness. No psychotropic medication was prescribed as part of the treatment, and any psychotropic drugs (ie, selective serotonin reuptake inhibitors) that were being taken at admission (by 8 patients) were gradually phased out during the first 2 wk of hospitalization.
Measurements All data were collected in the first week of inpatient admission and during the last week of the program.
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
hypothalamic-pituitary-adrenal axis (18), has been observed, but no data are yet available in patients with AN. Clearly, a confirmation of such an association would have key implications in the management of patients with AN and pave the way for clinical strategies to help them tolerate the physical changes they experience during recovery (16). In this context, as in other settings, dual-energy X-ray absorptiometry (DXA) is becoming an increasingly useful method of measuring body composition (19). However, previous DXA studies conducted on this issue have only examined certain patterns of body fat (ie, arms, legs, and trunk) with respect to total fat mass (ie, legs:total mass, arms:total mass, and trunk: total mass), and to our knowledge, no previous study has examined android or gynoid patterns in individuals with AN. Such an investigation would enable a more-accurate assessment of abdominal fat because android fat seems significantly associated with the clustering of metabolic syndrome components (20). It would also be of interest to compare such data with those of control subjects matched by age and patient posttreatment BMI. Therefore, we set out to 1) assess the quantity and distribution of lean and fat masses before and after complete weight restoration in patients with AN and 2) compare the body composition and fat distribution of weight-restored patients with those of healthy controls matched by age and patient posttreatment BMI. We also aimed to assess any relation between body-composition patterns and eating disorder psychopathology or general distress in patients with AN.
773
BODY COMPOSITION IN ANOREXIA NERVOSA
Body weight and height Body weight and height were measured by a medical doctor by using medical weighing scales and a stadiometer, respectively. Participants were weighed before breakfast wearing only underwear and no shoes. The BMI of patients was determined according to the standard formula of body weight measured in kilograms divided by height in meters squared. Body composition and body patterns
1) Total body composition, which included a) Total fat mass (kg) b) Fat mass as a percentage of the total mass (%), which was calculated as ½Total fat massOðtotal fat mass þ total lean mass þ total bone mineral content ½BMCÞ 3 100 ð1Þ c) Fat mass index (kg/m2) d) Total lean mass (kg) e) Lean mass as a percentage of total mass (%), which was calculated as ½Total lean massOðtotal fat mass þ total lean mass þ total BMCÞ 3 100 2
f) Lean mass index (kg/m ) 2) Regional body fat percentages, which included
½Arm fat massOðarm fat mass þ arm lean mass þ arm BMCÞ 3 100
ð3Þ
b) Percentage of leg region made up of fat (%), which was calculated as ½Leg fat massOðleg fat mass þ leg lean mass þ leg BMCÞ ð4Þ 3 100 c) Percentage of trunk region made up of fat (%), which was calculated as ½Trunk fat massOðtrunk fat mass þ trunk lean mass þ trunk BMCÞ 3 100
ð5Þ
d) Percentage of android region made up of fat (%), which was calculated as ½Android fat massO ðandroid fat mass þ android lean mass þ android BMCÞ 3 100 ð6Þ e) Percentage of gynoid region made up of fat (%), which was calculated as ½Gynoid fat massOðgynoid fat mass þ gynoid lean mass þgynoid BMCÞ 3 100
ð7Þ
3) Fat mass distribution, which included a) Android:gynoid ratio (ie, percentage fat in the android region:percentage of fat in the gynoid region) b) Arm fat as a percentage of total body fat (%), which was calculated as ðArm fat massOtotal fat massÞ 3 100
ð8Þ
c) Leg fat as a percentage of total body fat (%), which was calculated as ðLeg fat massOtotal fat massÞ 3 100
ð9Þ
d) Trunk fat as a percentage of total body fat (%), which was calculated as ðTrunk fat massOtotal fat massÞ 3 100
ð10Þ
ð2Þ Eating disorder psychopathology Eating disorder psychopathology was assessed by using the validated Italian version of the 12th edition of the EDE (21, 22).
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
DXA scans [Prodigy Primo Lunar, A223040501; General Electric Company; USA-enCORE TM 2009 (v13.31) software; GE Healthcare] were used to measure total and regional fat, lean, and bone masses. Scans were performed in the morning, and no special preparations were made except for instructing participants to wear only underwear and no metal accessories. Fat masses were measured in the arm, trunk, leg, and android and gynoid regions. The upper limit of the arm region was a hypothetical line from the crease of the axilla through the glenohumeral joint. The trunk region comprised the neck, chest, abdomen, and pelvis and extended from the inferior edge of the chin to a line that intersected the center of the femoral necks (below the pelvic brim). The leg region was taken to include all anatomic structures below this line. Within the trunk region, the android region was taken as the area between the ribs and the pelvis with an upper demarcation that was 20% of the distance between the iliac crest and neck and a lower perimeter that coincided with the top of the pelvis. The gynoid region, which, comprised the hips and upper thighs, overlapped parts of both the legs and trunk; the upper demarcation of this region fell below the top of the iliac crest at a distance 1.5 times the android height. The total height of the gynoid region was twice that of the android region. More details concerning regional body-composition analysis has been described elsewhere (25, 26). In this study, we assessed the following variables:
a) Percentage of arm region made up of fat (%), which was calculated as
774
EL GHOCH ET AL
General psychological distress General psychological distress was assessed by using a Global Severity Index (GSI) calculated from Brief Symptom Inventory (BSI) scores (validated Italian version) (30, 31). For the assessment, patients score 53 items, which are indicative of 9 symptom dimensions and 3 global indexes of distress, according to their level of concern regarding various symptoms over the preceding week [from 0 (not at all) to 4 (extremely concerned)]. The GSI is the single best indicator of current distress levels, and scores are calculated by summing and dividing by the total number of completed items. Reported norms for adult women indicate a mean (6SD) GSI of 0.35 6 0.37 for nonpatients (32). The BSI has been shown to have good internal consistency and reliability for the 9 dimensions, with a coefficients that ranged from 0.71 to 0.85 and test-retest reliability coefficients that ranged from 0.68 to 0.91. Good convergent, construct, and predictive validities have been reported (32). In our sample Cronbach’s a was 0.97.
from baseline to the end of treatment compared with end-oftreatment global EDE and shape-concern subscale scores, and compared with the BSI-GSI score; 5) between end-of-treatment body-composition variables compared with the change in global EDE and shape-concern subscale scores from baseline to the end of treatment, and compared with the BSI-GSI score; and 6) between the change between baseline and end-of-treatment body-composition variables compared with the change in global EDE and shape-concern subscale scores from baseline to the end of treatment, and compared with the BSI-GSI score. Significantly correlated variables were inserted into single linear regression models by using global EDE, shape-concern EDE subscale, or BSI-GSI scores as dependent variables, with age, educational level, physical activity level (measured in min/ mo according to EDE items), the presence of amenorrhea, and baseline questionnaire scores controlled for, as appropriate. To adjust P values for multiple comparisons, we applied Holm’s procedure, which is a less conservative and more powerful method than Bonferroni’s test (33, 34). P was significant at ,0.018. Data were analyzed with SPSS software (SPSS Statistics, version 20.0; IBM).
RESULTS
Baseline clinical characteristics of participants Body-composition variables of the 50 patients with AN and 100 controls are shown in Table 2. At baseline, patients with AN had significantly lower weight, BMI, total lean and fat masses, total fat mass percentage, fat and lean mass indexes, and a higher lean mass:fat mass ratio than those of controls. Arm, leg, trunk, and android and gynoid fat mass percentages were also significantly lower in patients with AN than controls, although there was no significant difference between the 2 groups in terms of the android:gynoid ratio. Finally, patients with AN displayed lower percentages of (arm fat O total fat) 3 100 and (trunk fat O total fat) 3 100 than those of controls but similar percentages of (leg fat O total fat) 3 100.
Statistical analysis
Effect of treatment
At baseline, demographic and clinical characteristics of both patients with AN and controls were tested for differences by using the t test for independent samples. The significance of differences in clinical variables from admission to discharge in patients with AN was determined by using the t test for paired data. To evaluate relations between the total and regional body composition and eating disorder psychopathology and general psychological distress, we applied Pearson’s product-moment correlation coefficients as follows: 1) between baseline bodycomposition variables compared with baseline global EDE and shape-concern subscale scores, and compared with the Global Severity Index of the Brief Symptom Inventory (BSI-GSI) score; 2) between end-of-treatment body-composition variables compared with end-of-treatment global EDE and shape-concern subscale scores, and compared with the BSI-GSI score; 3) between the change in body-composition variables from baseline to the end of treatment compared with baseline global EDE and shape-concern subscale scores, and compared with the BSI-GSI score; 4) between the change in body-composition variables
The treatment was associated with a significant increase in the mean (6SD) weight of patients (11.1 6 4.1 kg; P , 0.001), which was equivalent to a BMI increase of 4.3 6 1.6 (P , 0.001). Eating disorder psychopathology and general psychological distress also improved substantially by the program as evidenced by the mean global EDE score, which decreased by 1.7 6 1.3 (P , 0.001), and the mean BSI-GSI, which fell by 1.1 6 0.8 (P , 0.001). More than fifty percent of patients (54%; 27 of 50 women) had minimal residual eating disorder psychopathology measured as a global EDE score,1SD above the community mean (ie, ,1.74) (28), and 28% of patients (14 of 50 women) showed minimal residual general psychopathology (BSI-GSI ,0.35). A comparison of pretreatment and posttreatment body-composition variables in patients with AN indicated significant improvements in each variables from baseline to the end of treatment (Table 3). Percentages of (arm fat O total fat) 3 100 and (leg fat O total fat) 3 100 were not significantly different between baseline and the end of treatment, but the percentage of (trunk fatO total fat) 3 100
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
The EDE is an investigator-based interview used to assess the frequency of key behavioral and attitudinal aspects of eating disorders over the preceding 4 wk (28 d). The test is scored on a 7-point scale from 0 to 6, with higher scores reflecting greater severity or frequency. The EDE contains 4 subscales (dietary restraint, eating concern, weight concern, and shape concern) and a global score, which provides a detailed and comprehensive profile of the psychopathological features of eating disorder patients (27). The interrater reliability of the EDE has been estimated to be 0.97–0.99 (28), and in this case, the questionnaire was administered by assessors (trained and supervised by RDG, who is an expert in the use of the questionnaire). Assessors had no involvement in the treatment itself. EDE scores were used to generate 2 primary outcome variables as follows: the change in the severity of eating disorder features and possession of a global EDE score ,1 SD above the community mean (ie, ,1.74) (29). Finally, data were analyzed for any correlation between body-composition variables and global EDE or shape-concern subscale scores.
775
BODY COMPOSITION IN ANOREXIA NERVOSA TABLE 2 Body weight and body-composition variables in prettreatment and posttreatment patients with AN and the control group1 Comparison between posttreatment AN and control groups
Pretreatment AN (n = 50)
Posttreatment AN (n = 50)
Control group (n = 100)
Difference
P2
Difference
39.9 6 5.1 15.4 6 1.6
51.0 6 4.6 19.6 6 0.8
52.5 6 3.7 19.8 6 0.7
12.5 6 1.4 4.4 6 0.8
,0.001 ,0.001
1.5 6 0.8 0.2 6 0.1 0.4 0.2 0.07 0.9 0.003 0.07 20.2
6 6 6 6 6 6 6
1.0 1.8 0.4 0.7 0.06 0.3 0.7
Comparison between pretreatment and posttreatment AN values
P2
Change
P3
0.0374 0.219
11.1 6 4.1 4.3 6 1.6
,0.001 ,0.001
0.433 0.838 0.727 0.131 0.893 0.693 0.336
8.1 13.8 3.1 2.8 0.08 1.1 9.2
6 6 6 6 6 6 6
2.8 5.4 1.1 2.8 0.1 1.1 7.1
,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001
4.1 10.0 1.6 34.0 48.6 13.1 12.8
6 6 6 6 6 6 6
2.7 6.0 1.0 4.0 0.2 1.2 7.7
12.2 23.9 4.7 36.8 48.7 14.2 3.3
6 6 6 6 6 6 6
3.2 5.8 1.2 4.1 0.2 1.1 1.5
12.6 24.1 4.8 37.8 48.7 14.3 3.1
6 6 6 6 6 6 6
2.3 3.9 0.8 3.4 0.1 0.8 0.8
8.5 14.0 3.2 3.8 0.07 1.2 29.5
6 6 6 6 6 6 6
0.4 2.1 0.2 0.6 0.1 0.4 6.9
,0.001 ,0.001 ,0.001 ,0.001 0.0184 ,0.001 ,0.001
7.3 13.4 8.5 8.9 18.1
6 6 6 6 6
4.9 8.8 5.3 6.2 11.2
20.2 27.9 23.3 27.2 36.7
6 6 6 6 6
7.1 6.7 6.4 7.6 6.0
24.8 31.5 18.8 19.4 31.6
6 6 6 6 6
6.3 5.0 4.1 4.9 4.4
17.6 18.1 10.4 10.5 13.5
6 6 6 6 6
1.3 3.8 1.2 1.2 6.7
,0.001 ,0.001 ,0.001 ,0.001 ,0.001
4.7 3.6 4.5 7.8 5.1
6 6 6 6 6
0.8 1.8 2.2 2.6 1.6
,0.001 0.001 ,0.001 ,0.001 ,0.001
12.9 14.5 14.8 18.3 18.6
6 6 6 6 6
6.0 5.9 6.1 7.8 8.4
,0.001 ,0.001 ,0.001 ,0.001 ,0.001
0.6 6.5 43.7 40.8
6 6 6 6
0.2 1.9 10.2 7.7
0.7 7.0 42.5 46.1
6 6 6 6
0.1 1.6 6.3 6.3
0.6 10.5 47.3 35.7
6 6 6 6
0.1 1.6 4.8 4.3
0.02 4.1 3.5 25.1
6 6 6 6
0.1 0.3 5.4 3.5
0.508 ,0.001 0.0244 ,0.001
0.1 3.5 4.8 10.3
6 6 6 6
0.03 0.04 1.5 2.0
,0.001 ,0.001 ,0.001 ,0.001
0.1 0.6 21.3 5.3
6 6 6 6
0.2 1.9 7.3 5.5
,0.001 0.044 0.277 ,0.001
1 All values are means 6 SDs. Pretreatment assessment occurred in the first week of inpatient admission, and posttreatment assessment occurred in the last week of the 20-wk treatment program, which comprised 13 wk of inpatient therapy followed by 7 wk of partial hospitalization. AN, anorexia nervosa. 2 P values were determined by using the t test for independent samples. 3 P values were determined by using the t test for paired samples. 4 Not significant according to Holm’s procedure (P , 0.018).
was significantly higher after weight restoration than at baseline (Table 2). No significant differences were shown between control and posttreatment AN groups in terms of weight, BMI, and total, percentages, or indexes of lean and fat masses. However, after treatment, patients with AN had lower percentages of arm and leg fat masses; higher percentages of trunk, android, and gynoid fat masses; a higher android:gynoid ratio; lower percentages of (arm fat O total fat) 3 100 and (leg fat O total fat) 3 100; and
a higher percentage of (trunk fatO total fat) 3 100 than those in the control group (Table 2).
Body composition, eating disorder psychopathology, and general psychological distress No significant relation was shown between body-composition variables and global EDE or shape-concern EDE subscale scores, either at baseline, at the end of treatment, or in changes between
TABLE 3 Eating disorder psychopathology and psychological distress in pretreatment and posttreatment patients with AN1 Pretreatment AN (n = 50) Eating Disorder Examination (12th ed) Overall severity Dietary restraint Eating concern Weight concern Shape concern Brief Symptom Inventory Global Severity Index
3.4 3.8 3.3 3.1 3.6
6 6 6 6 6
1.4 1.6 1.6 1.9 1.6
1.9 6 0.9
Posttreatment AN (n = 50) 1.7 1.2 1.0 1.9 2.5
6 6 6 6 6
1.1 1.0 0.9 1.4 1.5
0.8 6 0.6
Change
P
6 6 6 6 6
1.3 1.6 1.6 1.7 1.7
,0.001 ,0.001 ,0.001 ,0.001 ,0.001
1.1 6 0.8
,0.001
1.7 2.5 2.3 1.2 1.1
All values are means 6 SDs. Pretreatment assessment occurred in the first week of inpatient admission, and posttreatment assessment occurred in the last week of the 20-wk treatment program, which comprised 13 wk of inpatient therapy followed by 7 wk of partial hospitalization. P values were determined by using the t test for paired samples. All P values were significant according to Holm’s procedure (P , 0.018). AN, anorexia nervosa. 1
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
Weight (kg) BMI (kg/m2) Total body composition Total fat mass (kg) Total fat mass (%) Fat mass index Total lean mass (kg) Total lean mass (%) Lean mass index Lean mass:fat mass ratio Regional body fat (%) Arm fat mass Leg fat mass Trunk fat mass Android fat mass Gynoid fat mass Fat mass distribution Android:gynoid ratio Arm fat O total fat (%) Leg fat O total fat (%) Trunk fat O total fat (%)
Comparison between pretreatment AN and control groups
776
EL GHOCH ET AL
DISCUSSION
Findings and concordance with previous studies The DXA assessment of body composition and fat distribution in 50 adult patients with AN before and after weight restoration and 100 healthy controls matched by age and patients’ posttreatment BMI revealed 3 main findings. The first finding was that, before treatment, patients with AN had lower BMI, lower total lean and fat masses, a lower total fat percentage, lower fat and lean indexes, and a higher lean mass:fat mass ratio than those of controls, but after weight restoration, no significant differences in these variables were shown. This result indicated that patients with AN restored their fat and lean masses completely during the 20 wk of the treatment administered. The second finding was that there was a marked difference in the body fat distribution between AN and control groups after weight restoration in the former group. Specifically, posttreatment patients with AN had higher fat percentages than did controls in central regions, especially in the android and gynoid regions, but lower fat percentages in peripheral regions. This result supported previous data that, during a relatively rapid weight gain, patients with AN tend to deposit fat mass prefer-
entially in the trunk region (8–11, 13, 35, 36). The android: gynoid ratio we measured also indicated that posttreatment patients with AN had a greater android fat deposition than that of controls. The greater accumulation of body fat in central regions, as opposed to extremities (ie, arm, forearm, thigh, and calf), could have explained the failure, reported in a previous study, of the skinfold technique to quantify the percentage of body fat in patients with AN after weight restoration (37). The third finding was the absence of any significant relation between eating disorder psychopathology and baseline or end-oftreatment body-composition variables. This result suggested that body shape did not influence the maintenance of the body image disturbance of patients as assessed by the EDE. Moreover, although changes in the BSI-GSI score and some body-composition variables were correlated, the linear regression analysis indicated that the change in the BSI-GSI score was significantly predicted only by the severity of psychological distress at baseline. Potential clinical implications The absence of any detectable relation between bodycomposition variables and eating disorder psychopathology should be discussed with patients. This discussion could help patients focus their efforts on addressing the psychological mechanisms that maintain their body-image disturbance rather than attempting to change the shape of their bodies. Likewise, it may be advisable for clinicians to shift the emphasis from changing the body shape and composition of patients with AN during nutritional rehabilitation to focusing more on addressing their eating disorder psychopathology. Moreover, clinicians should pay attention to the psychological distress of patients because this predicts the improvement in general psychopathology during treatment. Future research should be designed to explore clinical implications of increased truncal fat in women with AN. Indeed, in other populations (eg, the obese) an increase in body fat in the central regions (ie, trunk or android) is associated with metabolic syndrome and is a risk factor for cardiovascular disease, nonalcoholic fatty liver disease, insulin resistance (38), infertility, and ovulation disorders (39). Although one study showed that insulin-resistance onset was associated with a small increase in abdominal fat after a brief period of refeeding in patients with AN (40), no long-term effect data are available. Study strengths The study has several strengths. First, we assessed 3 levels of body composition [ie, total body composition, regional body fat percentages (which, to our knowledge, were not previously before assessed in patients with AN), and fat mass distribution] by using DXA software. DXA, although not yet accepted as a goldstandard technique for body-composition measurement, is known to exhibit a high level of precision in adults with AN, especially in the measurement of body fat (41). Furthermore, this method is rapid, relatively inexpensive, and only relies on a small amount of ionizing radiation. Moreover, the method enables the assessment of regional composition measurements, which provides a distinct advantage compared with approaches such as bioelectrical impedance or hydrodensitometry (42). Second, we assessed patients with AN both before and after body weight normalization, which
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
the 2 time points. However, significant correlations between the BSI-GSI score and several body-composition variables were detected (P , 0.05). In particular, the baseline BSI-GSI score was significantly correlated with baseline (arm fat O total fat) 3 100 (r = 20.31, P = 0.031). There were also significant correlations between the difference between baseline and end-oftreatment BSI-GSI scores and the following end-of-treatment body-composition variables: total fat mass percentage (r = 20.30, P = 0.036), total lean mass (r = 20.29, P = 0.040), lean mass:fat mass ratio (r = 20.31, P = 0.030), arm fat mass percentage (r = 20.40, P = 0.008), trunk fat mass percentage (r = 20.29, P = 0.044), android fat mass percentage (r = 20.32, P = 0.022), and (arm fat O total fat) 3 100 percentage (r = –0.39, P = 0.010). Finally, the change in the BSI-GSI score was significantly correlated with both the change in the arm fat mass percentage and change in the android fat mass percentage (r = 20.36, P = 0.011; r = 20.30, P = 0.032, respectively). Independent linear regression analyses were used to assess the relation between the change in the BSI-GSI score and bodycomposition variables, with age, education level, presence of amenorrhea, baseline BSI-GSI score, and physical activity level at the end of treatment controlled for. Results indicated that, for each independent variable, only the baseline BSI-GSI score was significantly related with the change in the BSI-GSI score [total fat mass percentage (b = –0.23, t = 21.73, P = 0.092), total lean mass (b = 0.19, t = 1.52, P = 0.135), ratio of lean mass:fat mass (b = 0.18, t = 1.46, P = 0.153), arm fat mass percentage (b = 20.24, t = 21.98, P = 0.055), trunk fat mass percentage (b = 20.16, t = 21.72, P = 0.093), android fat mass percentage (b = 20.24, t = 2197, P = 0.056), and (arm fat O total fat) 3 100 percentage (b = 20.12, t = 20.97, P = 0.338), change in arm fat mass percentage (b = 0.22, t = 1.90, P = 0.064), and change in android fat mass percentage (b = 0.20, t = 1.77; P = 0.085)], which indicated that the change in the BSI-GSI score was significantly predicted only by the severity of psychological distress at baseline.
BODY COMPOSITION IN ANOREXIA NERVOSA
enabled us to compare the 2 sets of data. Third, the EDE, which is the gold standard instrument for the assessment of eating disorder psychopathology, and BSI, which is widely recognized in clinical and research settings as a measure of psychological distress, were used. Fourth, our patients with AN were compared with a large sample of healthy controls matched by age and BMI achieved by patients with AN after treatment. Study limitations
The authors’ responsibilities were as follows—RDG and MEG: designed the study; MEG, CM, MP, and NCB: carried out the data assessment; SC: performed the statistical analysis; RDG, MEG, and SC: cowrote the manuscript; and all authors: claimed authorship and have approved and made substantial contributions to the conception, drafting, and final version of the manuscript. None of the authors had a conflict of interest.
REFERENCES 1. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 5th ed (DSM-5). Arlington, VA: American Psychiatric Publishing, 2013. 2. Dalle Grave R, Pasqualoni E, Marchesini G. Symptoms of starvation in eating disorder patients. In: Preedy VR, ed. Handbook of behavior, food and nutrition. New York: Springer Science+Business Media, 2011:2259–69. 3. Marzola E, Nasser JA, Hashim SA, Shih PA, Kaye WH. Nutritional rehabilitation in anorexia nervosa: review of the literature and implications for treatment. BMC Psychiatry 2013;13:290. 4. Fairburn CG. Cognitive behavior therapy and eating disorders. New York, NY: Guilford Press, 2008. 5. Frisch RE. Fatness and fertility. Sci Am 1988;258:88–95. 6. Krahn DD, Rock C, Dechert RE, Nairn KK, Hasse SA. Changes in resting energy expenditure and body composition in anorexia nervosa patients during refeeding. J Am Diet Assoc 1993;93:434–8. 7. Probst M, Goris M, Vandereycken W, Van Coppenolle H. Body composition in female anorexia nervosa patients. Br J Nutr 1996;76:639–47. 8. Iketani T, Kiriike N, Nagata T, Yamagami S. Altered body fat distribution after recovery of weight in patients with anorexia nervosa. Int J Eat Disord 1999;26:275–82. 9. Grinspoon S, Thomas L, Miller K, Pitts S, Herzog D, Klibanski A. Changes in regional fat redistribution and the effects of estrogen during spontaneous weight gain in women with anorexia nervosa. Am J Clin Nutr 2001;73:865–9. 10. Mayer L, Walsh BT, Pierson RN Jr, Heymsfield SB, Gallagher D, Wang J, Parides MK, Leibel RL, Warren MP, Killory E, et al. Body fat redistribution after weight gain in women with anorexia nervosa. Am J Clin Nutr 2005;81:1286–91. 11. Scalfi L, Polito A, Bianchi L, Marra M, Caldara A, Nicolai E, Contaldo F. Body composition changes in patients with anorexia nervosa after complete weight recovery. Eur J Clin Nutr 2002;56:15–20. 12. Tothill P, James Hannan W. Dual-energy X-ray absorptiometry measurements of fat and lean masses in subjects with eating disorders. Int J Obes Relat Metab Disord 2004;28:912–9. 13. Mayer LE, Klein DA, Black E, Attia E, Shen W, Mao X, Shungu DC, Punyanita M, Gallagher D, Wang J, et al. Adipose tissue distribution after weight restoration and weight maintenance in women with anorexia nervosa. Am J Clin Nutr 2009;90:1132–7. 14. Bodell LP, Mayer LE. Percent body fat is a risk factor for relapse in anorexia nervosa: a replication study. Int J Eat Disord 2011;44:118–23. 15. Mayer LE, Roberto CA, Glasofer DR, Etu SF, Gallagher D, Wang J, Heymsfield SB, Pierson RN Jr, Attia E, Devlin MJ, et al. Does percent body fat predict outcome in anorexia nervosa? Am J Psychiatry 2007; 164:970–2. 16. Mayer L. Body composition and anorexia nervosa: does physiology explain psychology? Am J Clin Nutr 2001;73:851–2. 17. Lee ES, Kim YH, Beck SH, Lee S, Oh SW. Depressive mood and abdominal fat distribution in overweight premenopausal women. Obes Res 2005;13:320–5. 18. Bornstein SR, Schuppenies A, Wong ML, Licinio J. Approaching the shared biology of obesity and depression: the stress axis as the locus of gene-environment interactions. Mol Psychiatry 2006;11:892–902. 19. Albanese CV, Diessel E, Genant HK. Clinical applications of body composition measurements using DXA. J Clin Densitom 2003;6:75– 85. 20. Kang SM, Yoon JW, Ahn HY, Kim SY, Lee KH, Shin H, Choi SH, Park KS, Jang HC, Lim S. Android fat depot is more closely associated with metabolic syndrome than abdominal visceral fat in elderly people. PLoS ONE 2011;6:e27694. 21. Mannucci E, Ricca V, Di Bernardo M, Rotella C. Studio del comportamento alimentare con una intervista strutturata: la Eating Disorder Examination. [Study of eating behavior with a structured interview: The Eating Disorder Examination.] Diabete 1996;8:127–31 (in Italian). 22. Fairburn CG, Cooper Z. The Eating Disorder Examination. 12th ed. In: Fairburn CG, Wilson GT, eds. Binge eating: nature, assessment, and treatment. New York, NY: Guilford Press, 1993:317–60. 23. Dalle Grave R. Multistep cognitive behavioral therapy for eating disorders: theory, practice, and clinical cases. New York, NY: Jason Aronson, 2013. 24. Istituto Nazionale di Ricerca per gli Alimenti e la Nutrizione (INRAN). Linee guida per una sana alimentazione italiana. [Italian Guidelines for
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
Our study did have some limitations. First and foremost, data were obtained in a single inpatient unit by applying one treatment program, which meant that external validation is required. Furthermore, our results need to be interpreted with caution because they may not apply to patients with AN treated in outpatient settings, where a slower rate of weight regain could result in a different regional fat accumulation. However, one outpatient study has already shown a central accumulation of body fat in these subjects, even though weight regain was slower (9). It should also be noted that the less intensive treatment offered in outpatient settings may have a less pronounced effect in terms of mitigating concerns and psychological distress of patients associated with central body fat accumulation. Second, DXA assessment may underestimate lean mass in underweight patients with AN because, as showed by Haas et al (43), patients who have a modest reduction in lean mass can be severely protein depleted. This effect is because starvationinduced changes in body composition may be associated with an increase in total body water, which can, in turn, mask low total body protein when a 3-compartment model, as in DXA, is used. Hence, direct methods may be more reliable means of assessing lean mass in underweight patients. Third, without a long-term follow-up, we were unable to clarify whether lipodystrophy characterized by the preferential accumulation of fat in the central region is a transitory phenomenon, as suggested by a study in 16 patients with AN who a maintained normal body weight 1 y after hospital discharge (13) or whether it persists long-term after weight restoration. Hence, future studies should also take into account the effect of time on body composition by serially assessing patients with AN and healthy subjects because healthy individuals tend to gain weight and abdominal fat as they age. Finally, we did not assess physical activity levels in the control group, which was a feature that could have affected weight and body composition as well as the interpretation of results. In conclusion, our data show that acute weight restoration is associated with a preferential distribution of body fat in central regions, which does not, however, seem to significantly influence eating disorder psychopathology or psychological distress in patients with AN. Nevertheless, future longitudinal studies are needed to clarify whether the preferential accumulation of body fat can influence AN relapse in the long term or the development of metabolic and cardiovascular complications.
777
778
25. 26. 27. 28. 29. 30.
32. 33. 34. 35.
healthy eating.] ed. 2003 (in Italian). Available from: http://www.inran. it/648/linee_guida.html. Stults-Kolehmainen MA, Stanforth PR, Bartholomew JB, Lu T, Abolt CJ, Sinha R. DXA estimates of fat in abdominal, trunk and hip regions varies by ethnicity in men. Nutr Diabetes 2013;3:e64. Mazess RB, Barden HS, Bisek JP, Hanson J. Dual-energy x-ray absorptiometry for total-body and regional bone-mineral and soft-tissue composition. Am J Clin Nutr 1990;51:1106–12. Cooper Z, Cooper PJ, Fairburn CG. The validity of the eating disorder examination and its subscales. Br J Psychiatry 1989;154:807–12. Wilson GT, Smith D. Assessment of bulimia nervosa: an evaluation of the Eating Disorders Examination. Int J Eat Disord 1989;8:173–9. Kendall PC, Marrs-Garcia A, Nath SR, Sheldrick RC. Normative comparisons for the evaluation of clinical significance. J Consult Clin Psychol 1999;67:285–99. De Leo D, Frisoni GB, Rozzini R, Trabucchi M. Italian community norms for the Brief Symptom Inventory in the elderly. Br J Clin Psychol 1993;32:209–13. Derogatis LR, Spencer PM. The Brief Symptom Inventory: administration, scoring, and procedures manual. Baltimore: Clinical Psychometric Research, 1982. Derogatis LR. BSI Brief Symptom Inventory. Administration, scoring, and procedures manual. 4th ed. Minneapolis, MN: National Computer Systems, 1993. Aickin M, Gensler H. Adjusting for multiple testing when reporting research results: the Bonferroni vs Holm methods. Am J Public Health 1996;86:726–8. Holm S. A simple sequentially rejective multiple test procedure. Scand J Stat 1979;6:65–70. Orphanidou CI, McCargar LJ, Birmingham CL, Belzberg AS. Changes in body composition and fat distribution after short-term weight gain in patients with anorexia nervosa. Am J Clin Nutr 1997;65:1034–41.
36. Zamboni M, Armellini F, Turcato E, Todisco P, Gallagher D, Dalle Grave R, Heymsfield S, Bosello O. Body fat distribution before and after weight gain in anorexia nervosa. Int J Obes Relat Metab Disord 1997;21:33–6. 37. El Ghoch M, Alberti M, Milanese C, Battistini NC, Pellegrini M, Capelli C, Calugi S, Dalle Grave R. Comparison between dual-energy X-ray absorptiometry and skinfolds thickness in assessing body fat in anorexia nervosa before and after weight restoration. Clin Nutr 2012; 31:911–6. 38. Carr MC, Brunzell JD. Abdominal obesity and dyslipidemia in the metabolic syndrome: importance of type 2 diabetes and familial combined hyperlipidemia in coronary artery disease risk. J Clin Endocrinol Metab 2004;89:2601–7. 39. Diamanti-Kandarakis E, Bergiele A. The influence of obesity on hyperandrogenism and infertility in the female. Obes Rev 2001;2:231–8. 40. Prioletta A, Muscogiuri G, Sorice GP, Lassandro AP, Mezza T, Policola C, Salomone E, Cipolla C, Della Casa S, Pontecorvi A, et al. In anorexia nervosa, even a small increase in abdominal fat is responsible for the appearance of insulin resistance. Clin Endocrinol (Oxf) 2011; 75:202–6. 41. Bredella MA, Ghomi RH, Thomas BJ, Torriani M, Brick DJ, Gerweck AV, Misra M, Klibanski A, Miller KK. Comparison of DXA and CT in the assessment of body composition in premenopausal women with obesity and anorexia nervosa. Obesity (Silver Spring) 2010;18:2227–33. 42. Ellis KJ. Human body composition: in vivo methods. Physiol Rev 2000;80:649–80. 43. Haas VK, Kohn MR, Clarke SD, Allen JR, Madden S, Mu¨ller MJ, Gaskin KJ. Body composition changes in female adolescents with anorexia nervosa. Am J Clin Nutr 2009;89:1005–10.
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
31.
EL GHOCH ET AL
INTRODUCTION
Anorexia nervosa (AN)4 is characterized by a restriction of energy intake relative to that required, which causes sufferers to become significantly underweight (1). Both the symptoms associated with low body weight and dietary restriction (starvation symptoms) and the psychopathology behind the eating disorder (2) can be addressed through nutritional rehabilitation (3), weight restoration, and cognitive behavioral therapy (4). These strategies are also necessary to restore an adequate quantity of body mass, both lean and fat, to enable physiologic processes (ie, a menstrual cycle and resting energy expenditure) to resume (5).
As to be expected from their undereating and underweight conditions, a marked reduction in both fat and lean masses in individuals with AN has been well documented (6, 7). Several studies have examined changes in body composition and body fat distribution in adults with AN. Of these studies, the majority of trials showed a preferential trunk fat accumulation associated with an acute weight regain immediately after treatment (8–11), although this was not a universal finding (12). Longer term, one study showed no differences in visceral adipose tissue distribution between former patients with AN who had maintained normal weight for 1 y and BMI-matched control women with no history of AN (13). However, a lower percentage of body fat restoration after weight restoration does seem to be a risk factor for relapse in patients with AN (14, 15). Nevertheless, the research conducted to date on this issue has suffered from various methodologic limitations, such as small sample groups (8, 9), incomplete weight restoration (8, 9), and small (8, 10, 11, 13) and unsuitable control groups (normalweight but not lean participants) (12). Thus far, research has also failed to address the relation between fat distribution, eating disorder psychopathology, and the psychological distress associated with weight restoration in AN (16). Indeed, although some authors have hypothesized that patients with greater trunk fat deposition after weight restoration could experience considerable stress brought on by changes in their body shapes, which make them more prone to a relapse (10), data to support this theory are lacking. A reported association between depressive mood and visceral adipose tissue in overweight women (17), possibly involving the leptin-mediated dysregulation of the
1 From the Department of Eating and Weight Disorders, Villa Garda Hospital, Garda, Italy (MEG, SC, and RDG); the Department of Neurological and Movement Sciences, University of Verona, Verona, Italy (CM); and the Department of Diagnostic, Clinical and Public Health, University of Modena and Reggio Emilia, Modena and Reggio Emilia, Italy (MP and NCB). 2 No support was received for this study. 3 Address correspondence to M El Ghoch, Department of Eating and Weight Disorders, Villa Garda Hospital, Via Montebaldo, 89, I-37016 Garda, Verona, Italy. E-mail: [email protected]. 4 Abbreviations used: AN, anorexia nervosa; BMC, bone mineral content; BSI, Brief Symptom Inventory; BSI-GSI, Global Severity Index of the Brief Symptom Inventory; CBT-E, enhanced form of cognitive behavioral therapy; DXA, dual-energy X-ray absorptiometry; EDE, Eating Disorder Examination interview; GSI, Global Severity Index. Received October 25, 2013. Accepted for publication January 21, 2014. First published online February 5, 2014; doi: 10.3945/ajcn.113.078816.
Am J Clin Nutr 2014;99:771–8. Printed in USA. Ó 2014 American Society for Nutrition
771
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
ABSTRACT Background: Although the effect of immediate weight restoration on body composition and body fat distribution has previously been studied in anorexia nervosa (AN), its influence in women with AN on eating disorder psychopathology and psychological distress has not previously been investigated to our knowledge. Objectives: We assessed body composition and fat mass distribution before and after body weight restoration and investigated any relation between changes in body fat patterns of patients with AN treated in a specialist inpatient unit and their eating disorder and psychological distress features. Design: Body composition was measured by using dual-energy X-ray absorptiometry in 50 female, adult patients with AN before and after complete weight restoration [body mass index (BMI; in kg/m2) $18.5] and 100 healthy control subjects matched by age and posttreatment BMI of study group participants. Eating disorder psychopathology and psychological distress were assessed in the AN group before and after weight restoration by using the Eating Disorder Examination interview and the Global Severity Index of the Brief Symptom Inventory (BSI-GSI), respectively. Results: After the achievement of complete weight restoration, patients with AN had higher trunk (P , 0.001), android (P , 0.001), and gynoid (P , 0.001) fat masses and lower arm (P , 0.001) and leg (P = 0.001) fat masses with respect to control subjects. No relation was shown between body-composition variables and eating disorder psychopathology in the AN group, and the only significant predictor of change in BSI-GSI was the baseline BSI-GSI score. Conclusion: The normalization of body weight in patients with AN is associated with a preferential distribution of body fat in central regions, which does not, however, seem to influence either eating disorder psychopathology or psychological distress scores. This trial was registered at www.controlled-trials.com as ISRCTN92626057. Am J Clin Nutr 2014;99:771–8.
772
EL GHOCH ET AL
SUBJECTS AND METHODS
Participants Fifty white, female patients with AN and 100 normal-weight, white, female subjects were enrolled in the study. All 50 patients were voluntarily and consecutively admitted to the eating disorder inpatient unit of Villa Garda Hospital during the years 2011–2013. Inclusion criteria were as follows: 1) aged 18–50 y, 2) BMI (in kg/m2) #17.5, 3) Eating Disorder Examination interview (EDE; 12.0D) diagnosis of AN except the amenorrhea criterion (21, 22), 4) either failure of less-intensive outpatient treatment or an eating disorder of clinical severity that was not manageable in an outpatient setting, and 5) BMI $18.5 at the end of the 20-wk treatment. Patients with active substance abuse, schizophrenia, and other psychotic disorders were excluded. Demographic and baseline clinical characteristics of patients with AN are presented in Table 1. Patients with AN were carefully matched with 100 control individuals of the same age (62 y) with a BMI equivalent to patients’ BMI at the end of their treatment. Control subjects were recruited randomly from university listservs and advertisements and were all healthy and weight-stable, with regular menstruation, no history of eating disorders or other significant psychiatric or medical conditions, and no current psychotropic medication intake. Inclusion criteria for both patients and controls were evaluated during an eligibility interview conducted by a specialist in the field (RDG). The study design was reviewed and approved by the Institutional Review Board of Villa Garda Hospital, Verona, and all participants gave written informed consent for the use of their anonymous personal data.
TABLE 1 Demographic and baseline clinical characteristics in patients with AN and the control group1
Age (y) Highest educational qualification [n (%)] Elementary school Middle school certificate High school diploma Bachelor’s degree Missing data Presence of amenorrhea [n (%)] Duration of illness (y)3 Amount of exercise (min)3,4
Pretreatment AN (n = 50)
Control group (n = 100)
26.5 6 7.62
26.5 6 7.6
0 13 (26.0) 26 (52.0) 9 (18.0) 2 (4.0) 42 (84.0) 8.7 6 6.7 1533.9 6 2732.3
0 0 50 (50.0) 50 (50.0) 0 0 (0.0) — —
1
AN, anorexia nervosa. Mean 6 SD (all such values). 3 Data were not available for the control group. 4 In the past 28 d (measured by using the Eating Disorder Examination interview, 12th ed). 2
Inpatient treatment protocol The inpatient treatment has been described in detail elsewhere (23). In brief, the treatment was an adapted version of the enhanced form of cognitive behavioral therapy (CBT-E) for eating disorders (4) that lasts 20 wk and comprises 13 wk of inpatient therapy followed by 7 wk of partial hospitalization. The program retained all main strategies and procedures of the CBT-E, which are delivered in both individual and group sessions. However, the program differed from the outpatient CBT-E in that patients received dietitian-assisted eating in the early weeks of treatment until their BMI reached $18.5. Over the first 3 wk of treatment, the dietary intake of patients was increased incrementally from 1500 to 2500 kcal. In subsequent weeks of the program, the dietary energy content was adjusted to maintain a weight gain rate from 1 to 1.5 kg/wk until the BMI of patients reached $19. Once this target was achieved, the calorie content of meals was continually adjusted so that body weight of patients remained within a 2-kg range of this figure. The diet included all main food groups and conformed to Italian National Guidelines for Healthy Eating (24). In addition, patients with BMI $15 and stable medical conditions participated in physiotherapist-led 60-min physical exercise sessions 2 times/wk as part of their treatment. These sessions included calisthenic exercises to help restore muscle strength, flexibility, and posture as well as aerobic exercises to improve cardiovascular fitness. No psychotropic medication was prescribed as part of the treatment, and any psychotropic drugs (ie, selective serotonin reuptake inhibitors) that were being taken at admission (by 8 patients) were gradually phased out during the first 2 wk of hospitalization.
Measurements All data were collected in the first week of inpatient admission and during the last week of the program.
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
hypothalamic-pituitary-adrenal axis (18), has been observed, but no data are yet available in patients with AN. Clearly, a confirmation of such an association would have key implications in the management of patients with AN and pave the way for clinical strategies to help them tolerate the physical changes they experience during recovery (16). In this context, as in other settings, dual-energy X-ray absorptiometry (DXA) is becoming an increasingly useful method of measuring body composition (19). However, previous DXA studies conducted on this issue have only examined certain patterns of body fat (ie, arms, legs, and trunk) with respect to total fat mass (ie, legs:total mass, arms:total mass, and trunk: total mass), and to our knowledge, no previous study has examined android or gynoid patterns in individuals with AN. Such an investigation would enable a more-accurate assessment of abdominal fat because android fat seems significantly associated with the clustering of metabolic syndrome components (20). It would also be of interest to compare such data with those of control subjects matched by age and patient posttreatment BMI. Therefore, we set out to 1) assess the quantity and distribution of lean and fat masses before and after complete weight restoration in patients with AN and 2) compare the body composition and fat distribution of weight-restored patients with those of healthy controls matched by age and patient posttreatment BMI. We also aimed to assess any relation between body-composition patterns and eating disorder psychopathology or general distress in patients with AN.
773
BODY COMPOSITION IN ANOREXIA NERVOSA
Body weight and height Body weight and height were measured by a medical doctor by using medical weighing scales and a stadiometer, respectively. Participants were weighed before breakfast wearing only underwear and no shoes. The BMI of patients was determined according to the standard formula of body weight measured in kilograms divided by height in meters squared. Body composition and body patterns
1) Total body composition, which included a) Total fat mass (kg) b) Fat mass as a percentage of the total mass (%), which was calculated as ½Total fat massOðtotal fat mass þ total lean mass þ total bone mineral content ½BMCÞ 3 100 ð1Þ c) Fat mass index (kg/m2) d) Total lean mass (kg) e) Lean mass as a percentage of total mass (%), which was calculated as ½Total lean massOðtotal fat mass þ total lean mass þ total BMCÞ 3 100 2
f) Lean mass index (kg/m ) 2) Regional body fat percentages, which included
½Arm fat massOðarm fat mass þ arm lean mass þ arm BMCÞ 3 100
ð3Þ
b) Percentage of leg region made up of fat (%), which was calculated as ½Leg fat massOðleg fat mass þ leg lean mass þ leg BMCÞ ð4Þ 3 100 c) Percentage of trunk region made up of fat (%), which was calculated as ½Trunk fat massOðtrunk fat mass þ trunk lean mass þ trunk BMCÞ 3 100
ð5Þ
d) Percentage of android region made up of fat (%), which was calculated as ½Android fat massO ðandroid fat mass þ android lean mass þ android BMCÞ 3 100 ð6Þ e) Percentage of gynoid region made up of fat (%), which was calculated as ½Gynoid fat massOðgynoid fat mass þ gynoid lean mass þgynoid BMCÞ 3 100
ð7Þ
3) Fat mass distribution, which included a) Android:gynoid ratio (ie, percentage fat in the android region:percentage of fat in the gynoid region) b) Arm fat as a percentage of total body fat (%), which was calculated as ðArm fat massOtotal fat massÞ 3 100
ð8Þ
c) Leg fat as a percentage of total body fat (%), which was calculated as ðLeg fat massOtotal fat massÞ 3 100
ð9Þ
d) Trunk fat as a percentage of total body fat (%), which was calculated as ðTrunk fat massOtotal fat massÞ 3 100
ð10Þ
ð2Þ Eating disorder psychopathology Eating disorder psychopathology was assessed by using the validated Italian version of the 12th edition of the EDE (21, 22).
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
DXA scans [Prodigy Primo Lunar, A223040501; General Electric Company; USA-enCORE TM 2009 (v13.31) software; GE Healthcare] were used to measure total and regional fat, lean, and bone masses. Scans were performed in the morning, and no special preparations were made except for instructing participants to wear only underwear and no metal accessories. Fat masses were measured in the arm, trunk, leg, and android and gynoid regions. The upper limit of the arm region was a hypothetical line from the crease of the axilla through the glenohumeral joint. The trunk region comprised the neck, chest, abdomen, and pelvis and extended from the inferior edge of the chin to a line that intersected the center of the femoral necks (below the pelvic brim). The leg region was taken to include all anatomic structures below this line. Within the trunk region, the android region was taken as the area between the ribs and the pelvis with an upper demarcation that was 20% of the distance between the iliac crest and neck and a lower perimeter that coincided with the top of the pelvis. The gynoid region, which, comprised the hips and upper thighs, overlapped parts of both the legs and trunk; the upper demarcation of this region fell below the top of the iliac crest at a distance 1.5 times the android height. The total height of the gynoid region was twice that of the android region. More details concerning regional body-composition analysis has been described elsewhere (25, 26). In this study, we assessed the following variables:
a) Percentage of arm region made up of fat (%), which was calculated as
774
EL GHOCH ET AL
General psychological distress General psychological distress was assessed by using a Global Severity Index (GSI) calculated from Brief Symptom Inventory (BSI) scores (validated Italian version) (30, 31). For the assessment, patients score 53 items, which are indicative of 9 symptom dimensions and 3 global indexes of distress, according to their level of concern regarding various symptoms over the preceding week [from 0 (not at all) to 4 (extremely concerned)]. The GSI is the single best indicator of current distress levels, and scores are calculated by summing and dividing by the total number of completed items. Reported norms for adult women indicate a mean (6SD) GSI of 0.35 6 0.37 for nonpatients (32). The BSI has been shown to have good internal consistency and reliability for the 9 dimensions, with a coefficients that ranged from 0.71 to 0.85 and test-retest reliability coefficients that ranged from 0.68 to 0.91. Good convergent, construct, and predictive validities have been reported (32). In our sample Cronbach’s a was 0.97.
from baseline to the end of treatment compared with end-oftreatment global EDE and shape-concern subscale scores, and compared with the BSI-GSI score; 5) between end-of-treatment body-composition variables compared with the change in global EDE and shape-concern subscale scores from baseline to the end of treatment, and compared with the BSI-GSI score; and 6) between the change between baseline and end-of-treatment body-composition variables compared with the change in global EDE and shape-concern subscale scores from baseline to the end of treatment, and compared with the BSI-GSI score. Significantly correlated variables were inserted into single linear regression models by using global EDE, shape-concern EDE subscale, or BSI-GSI scores as dependent variables, with age, educational level, physical activity level (measured in min/ mo according to EDE items), the presence of amenorrhea, and baseline questionnaire scores controlled for, as appropriate. To adjust P values for multiple comparisons, we applied Holm’s procedure, which is a less conservative and more powerful method than Bonferroni’s test (33, 34). P was significant at ,0.018. Data were analyzed with SPSS software (SPSS Statistics, version 20.0; IBM).
RESULTS
Baseline clinical characteristics of participants Body-composition variables of the 50 patients with AN and 100 controls are shown in Table 2. At baseline, patients with AN had significantly lower weight, BMI, total lean and fat masses, total fat mass percentage, fat and lean mass indexes, and a higher lean mass:fat mass ratio than those of controls. Arm, leg, trunk, and android and gynoid fat mass percentages were also significantly lower in patients with AN than controls, although there was no significant difference between the 2 groups in terms of the android:gynoid ratio. Finally, patients with AN displayed lower percentages of (arm fat O total fat) 3 100 and (trunk fat O total fat) 3 100 than those of controls but similar percentages of (leg fat O total fat) 3 100.
Statistical analysis
Effect of treatment
At baseline, demographic and clinical characteristics of both patients with AN and controls were tested for differences by using the t test for independent samples. The significance of differences in clinical variables from admission to discharge in patients with AN was determined by using the t test for paired data. To evaluate relations between the total and regional body composition and eating disorder psychopathology and general psychological distress, we applied Pearson’s product-moment correlation coefficients as follows: 1) between baseline bodycomposition variables compared with baseline global EDE and shape-concern subscale scores, and compared with the Global Severity Index of the Brief Symptom Inventory (BSI-GSI) score; 2) between end-of-treatment body-composition variables compared with end-of-treatment global EDE and shape-concern subscale scores, and compared with the BSI-GSI score; 3) between the change in body-composition variables from baseline to the end of treatment compared with baseline global EDE and shape-concern subscale scores, and compared with the BSI-GSI score; 4) between the change in body-composition variables
The treatment was associated with a significant increase in the mean (6SD) weight of patients (11.1 6 4.1 kg; P , 0.001), which was equivalent to a BMI increase of 4.3 6 1.6 (P , 0.001). Eating disorder psychopathology and general psychological distress also improved substantially by the program as evidenced by the mean global EDE score, which decreased by 1.7 6 1.3 (P , 0.001), and the mean BSI-GSI, which fell by 1.1 6 0.8 (P , 0.001). More than fifty percent of patients (54%; 27 of 50 women) had minimal residual eating disorder psychopathology measured as a global EDE score,1SD above the community mean (ie, ,1.74) (28), and 28% of patients (14 of 50 women) showed minimal residual general psychopathology (BSI-GSI ,0.35). A comparison of pretreatment and posttreatment body-composition variables in patients with AN indicated significant improvements in each variables from baseline to the end of treatment (Table 3). Percentages of (arm fat O total fat) 3 100 and (leg fat O total fat) 3 100 were not significantly different between baseline and the end of treatment, but the percentage of (trunk fatO total fat) 3 100
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
The EDE is an investigator-based interview used to assess the frequency of key behavioral and attitudinal aspects of eating disorders over the preceding 4 wk (28 d). The test is scored on a 7-point scale from 0 to 6, with higher scores reflecting greater severity or frequency. The EDE contains 4 subscales (dietary restraint, eating concern, weight concern, and shape concern) and a global score, which provides a detailed and comprehensive profile of the psychopathological features of eating disorder patients (27). The interrater reliability of the EDE has been estimated to be 0.97–0.99 (28), and in this case, the questionnaire was administered by assessors (trained and supervised by RDG, who is an expert in the use of the questionnaire). Assessors had no involvement in the treatment itself. EDE scores were used to generate 2 primary outcome variables as follows: the change in the severity of eating disorder features and possession of a global EDE score ,1 SD above the community mean (ie, ,1.74) (29). Finally, data were analyzed for any correlation between body-composition variables and global EDE or shape-concern subscale scores.
775
BODY COMPOSITION IN ANOREXIA NERVOSA TABLE 2 Body weight and body-composition variables in prettreatment and posttreatment patients with AN and the control group1 Comparison between posttreatment AN and control groups
Pretreatment AN (n = 50)
Posttreatment AN (n = 50)
Control group (n = 100)
Difference
P2
Difference
39.9 6 5.1 15.4 6 1.6
51.0 6 4.6 19.6 6 0.8
52.5 6 3.7 19.8 6 0.7
12.5 6 1.4 4.4 6 0.8
,0.001 ,0.001
1.5 6 0.8 0.2 6 0.1 0.4 0.2 0.07 0.9 0.003 0.07 20.2
6 6 6 6 6 6 6
1.0 1.8 0.4 0.7 0.06 0.3 0.7
Comparison between pretreatment and posttreatment AN values
P2
Change
P3
0.0374 0.219
11.1 6 4.1 4.3 6 1.6
,0.001 ,0.001
0.433 0.838 0.727 0.131 0.893 0.693 0.336
8.1 13.8 3.1 2.8 0.08 1.1 9.2
6 6 6 6 6 6 6
2.8 5.4 1.1 2.8 0.1 1.1 7.1
,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001
4.1 10.0 1.6 34.0 48.6 13.1 12.8
6 6 6 6 6 6 6
2.7 6.0 1.0 4.0 0.2 1.2 7.7
12.2 23.9 4.7 36.8 48.7 14.2 3.3
6 6 6 6 6 6 6
3.2 5.8 1.2 4.1 0.2 1.1 1.5
12.6 24.1 4.8 37.8 48.7 14.3 3.1
6 6 6 6 6 6 6
2.3 3.9 0.8 3.4 0.1 0.8 0.8
8.5 14.0 3.2 3.8 0.07 1.2 29.5
6 6 6 6 6 6 6
0.4 2.1 0.2 0.6 0.1 0.4 6.9
,0.001 ,0.001 ,0.001 ,0.001 0.0184 ,0.001 ,0.001
7.3 13.4 8.5 8.9 18.1
6 6 6 6 6
4.9 8.8 5.3 6.2 11.2
20.2 27.9 23.3 27.2 36.7
6 6 6 6 6
7.1 6.7 6.4 7.6 6.0
24.8 31.5 18.8 19.4 31.6
6 6 6 6 6
6.3 5.0 4.1 4.9 4.4
17.6 18.1 10.4 10.5 13.5
6 6 6 6 6
1.3 3.8 1.2 1.2 6.7
,0.001 ,0.001 ,0.001 ,0.001 ,0.001
4.7 3.6 4.5 7.8 5.1
6 6 6 6 6
0.8 1.8 2.2 2.6 1.6
,0.001 0.001 ,0.001 ,0.001 ,0.001
12.9 14.5 14.8 18.3 18.6
6 6 6 6 6
6.0 5.9 6.1 7.8 8.4
,0.001 ,0.001 ,0.001 ,0.001 ,0.001
0.6 6.5 43.7 40.8
6 6 6 6
0.2 1.9 10.2 7.7
0.7 7.0 42.5 46.1
6 6 6 6
0.1 1.6 6.3 6.3
0.6 10.5 47.3 35.7
6 6 6 6
0.1 1.6 4.8 4.3
0.02 4.1 3.5 25.1
6 6 6 6
0.1 0.3 5.4 3.5
0.508 ,0.001 0.0244 ,0.001
0.1 3.5 4.8 10.3
6 6 6 6
0.03 0.04 1.5 2.0
,0.001 ,0.001 ,0.001 ,0.001
0.1 0.6 21.3 5.3
6 6 6 6
0.2 1.9 7.3 5.5
,0.001 0.044 0.277 ,0.001
1 All values are means 6 SDs. Pretreatment assessment occurred in the first week of inpatient admission, and posttreatment assessment occurred in the last week of the 20-wk treatment program, which comprised 13 wk of inpatient therapy followed by 7 wk of partial hospitalization. AN, anorexia nervosa. 2 P values were determined by using the t test for independent samples. 3 P values were determined by using the t test for paired samples. 4 Not significant according to Holm’s procedure (P , 0.018).
was significantly higher after weight restoration than at baseline (Table 2). No significant differences were shown between control and posttreatment AN groups in terms of weight, BMI, and total, percentages, or indexes of lean and fat masses. However, after treatment, patients with AN had lower percentages of arm and leg fat masses; higher percentages of trunk, android, and gynoid fat masses; a higher android:gynoid ratio; lower percentages of (arm fat O total fat) 3 100 and (leg fat O total fat) 3 100; and
a higher percentage of (trunk fatO total fat) 3 100 than those in the control group (Table 2).
Body composition, eating disorder psychopathology, and general psychological distress No significant relation was shown between body-composition variables and global EDE or shape-concern EDE subscale scores, either at baseline, at the end of treatment, or in changes between
TABLE 3 Eating disorder psychopathology and psychological distress in pretreatment and posttreatment patients with AN1 Pretreatment AN (n = 50) Eating Disorder Examination (12th ed) Overall severity Dietary restraint Eating concern Weight concern Shape concern Brief Symptom Inventory Global Severity Index
3.4 3.8 3.3 3.1 3.6
6 6 6 6 6
1.4 1.6 1.6 1.9 1.6
1.9 6 0.9
Posttreatment AN (n = 50) 1.7 1.2 1.0 1.9 2.5
6 6 6 6 6
1.1 1.0 0.9 1.4 1.5
0.8 6 0.6
Change
P
6 6 6 6 6
1.3 1.6 1.6 1.7 1.7
,0.001 ,0.001 ,0.001 ,0.001 ,0.001
1.1 6 0.8
,0.001
1.7 2.5 2.3 1.2 1.1
All values are means 6 SDs. Pretreatment assessment occurred in the first week of inpatient admission, and posttreatment assessment occurred in the last week of the 20-wk treatment program, which comprised 13 wk of inpatient therapy followed by 7 wk of partial hospitalization. P values were determined by using the t test for paired samples. All P values were significant according to Holm’s procedure (P , 0.018). AN, anorexia nervosa. 1
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
Weight (kg) BMI (kg/m2) Total body composition Total fat mass (kg) Total fat mass (%) Fat mass index Total lean mass (kg) Total lean mass (%) Lean mass index Lean mass:fat mass ratio Regional body fat (%) Arm fat mass Leg fat mass Trunk fat mass Android fat mass Gynoid fat mass Fat mass distribution Android:gynoid ratio Arm fat O total fat (%) Leg fat O total fat (%) Trunk fat O total fat (%)
Comparison between pretreatment AN and control groups
776
EL GHOCH ET AL
DISCUSSION
Findings and concordance with previous studies The DXA assessment of body composition and fat distribution in 50 adult patients with AN before and after weight restoration and 100 healthy controls matched by age and patients’ posttreatment BMI revealed 3 main findings. The first finding was that, before treatment, patients with AN had lower BMI, lower total lean and fat masses, a lower total fat percentage, lower fat and lean indexes, and a higher lean mass:fat mass ratio than those of controls, but after weight restoration, no significant differences in these variables were shown. This result indicated that patients with AN restored their fat and lean masses completely during the 20 wk of the treatment administered. The second finding was that there was a marked difference in the body fat distribution between AN and control groups after weight restoration in the former group. Specifically, posttreatment patients with AN had higher fat percentages than did controls in central regions, especially in the android and gynoid regions, but lower fat percentages in peripheral regions. This result supported previous data that, during a relatively rapid weight gain, patients with AN tend to deposit fat mass prefer-
entially in the trunk region (8–11, 13, 35, 36). The android: gynoid ratio we measured also indicated that posttreatment patients with AN had a greater android fat deposition than that of controls. The greater accumulation of body fat in central regions, as opposed to extremities (ie, arm, forearm, thigh, and calf), could have explained the failure, reported in a previous study, of the skinfold technique to quantify the percentage of body fat in patients with AN after weight restoration (37). The third finding was the absence of any significant relation between eating disorder psychopathology and baseline or end-oftreatment body-composition variables. This result suggested that body shape did not influence the maintenance of the body image disturbance of patients as assessed by the EDE. Moreover, although changes in the BSI-GSI score and some body-composition variables were correlated, the linear regression analysis indicated that the change in the BSI-GSI score was significantly predicted only by the severity of psychological distress at baseline. Potential clinical implications The absence of any detectable relation between bodycomposition variables and eating disorder psychopathology should be discussed with patients. This discussion could help patients focus their efforts on addressing the psychological mechanisms that maintain their body-image disturbance rather than attempting to change the shape of their bodies. Likewise, it may be advisable for clinicians to shift the emphasis from changing the body shape and composition of patients with AN during nutritional rehabilitation to focusing more on addressing their eating disorder psychopathology. Moreover, clinicians should pay attention to the psychological distress of patients because this predicts the improvement in general psychopathology during treatment. Future research should be designed to explore clinical implications of increased truncal fat in women with AN. Indeed, in other populations (eg, the obese) an increase in body fat in the central regions (ie, trunk or android) is associated with metabolic syndrome and is a risk factor for cardiovascular disease, nonalcoholic fatty liver disease, insulin resistance (38), infertility, and ovulation disorders (39). Although one study showed that insulin-resistance onset was associated with a small increase in abdominal fat after a brief period of refeeding in patients with AN (40), no long-term effect data are available. Study strengths The study has several strengths. First, we assessed 3 levels of body composition [ie, total body composition, regional body fat percentages (which, to our knowledge, were not previously before assessed in patients with AN), and fat mass distribution] by using DXA software. DXA, although not yet accepted as a goldstandard technique for body-composition measurement, is known to exhibit a high level of precision in adults with AN, especially in the measurement of body fat (41). Furthermore, this method is rapid, relatively inexpensive, and only relies on a small amount of ionizing radiation. Moreover, the method enables the assessment of regional composition measurements, which provides a distinct advantage compared with approaches such as bioelectrical impedance or hydrodensitometry (42). Second, we assessed patients with AN both before and after body weight normalization, which
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
the 2 time points. However, significant correlations between the BSI-GSI score and several body-composition variables were detected (P , 0.05). In particular, the baseline BSI-GSI score was significantly correlated with baseline (arm fat O total fat) 3 100 (r = 20.31, P = 0.031). There were also significant correlations between the difference between baseline and end-oftreatment BSI-GSI scores and the following end-of-treatment body-composition variables: total fat mass percentage (r = 20.30, P = 0.036), total lean mass (r = 20.29, P = 0.040), lean mass:fat mass ratio (r = 20.31, P = 0.030), arm fat mass percentage (r = 20.40, P = 0.008), trunk fat mass percentage (r = 20.29, P = 0.044), android fat mass percentage (r = 20.32, P = 0.022), and (arm fat O total fat) 3 100 percentage (r = –0.39, P = 0.010). Finally, the change in the BSI-GSI score was significantly correlated with both the change in the arm fat mass percentage and change in the android fat mass percentage (r = 20.36, P = 0.011; r = 20.30, P = 0.032, respectively). Independent linear regression analyses were used to assess the relation between the change in the BSI-GSI score and bodycomposition variables, with age, education level, presence of amenorrhea, baseline BSI-GSI score, and physical activity level at the end of treatment controlled for. Results indicated that, for each independent variable, only the baseline BSI-GSI score was significantly related with the change in the BSI-GSI score [total fat mass percentage (b = –0.23, t = 21.73, P = 0.092), total lean mass (b = 0.19, t = 1.52, P = 0.135), ratio of lean mass:fat mass (b = 0.18, t = 1.46, P = 0.153), arm fat mass percentage (b = 20.24, t = 21.98, P = 0.055), trunk fat mass percentage (b = 20.16, t = 21.72, P = 0.093), android fat mass percentage (b = 20.24, t = 2197, P = 0.056), and (arm fat O total fat) 3 100 percentage (b = 20.12, t = 20.97, P = 0.338), change in arm fat mass percentage (b = 0.22, t = 1.90, P = 0.064), and change in android fat mass percentage (b = 0.20, t = 1.77; P = 0.085)], which indicated that the change in the BSI-GSI score was significantly predicted only by the severity of psychological distress at baseline.
BODY COMPOSITION IN ANOREXIA NERVOSA
enabled us to compare the 2 sets of data. Third, the EDE, which is the gold standard instrument for the assessment of eating disorder psychopathology, and BSI, which is widely recognized in clinical and research settings as a measure of psychological distress, were used. Fourth, our patients with AN were compared with a large sample of healthy controls matched by age and BMI achieved by patients with AN after treatment. Study limitations
The authors’ responsibilities were as follows—RDG and MEG: designed the study; MEG, CM, MP, and NCB: carried out the data assessment; SC: performed the statistical analysis; RDG, MEG, and SC: cowrote the manuscript; and all authors: claimed authorship and have approved and made substantial contributions to the conception, drafting, and final version of the manuscript. None of the authors had a conflict of interest.
REFERENCES 1. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 5th ed (DSM-5). Arlington, VA: American Psychiatric Publishing, 2013. 2. Dalle Grave R, Pasqualoni E, Marchesini G. Symptoms of starvation in eating disorder patients. In: Preedy VR, ed. Handbook of behavior, food and nutrition. New York: Springer Science+Business Media, 2011:2259–69. 3. Marzola E, Nasser JA, Hashim SA, Shih PA, Kaye WH. Nutritional rehabilitation in anorexia nervosa: review of the literature and implications for treatment. BMC Psychiatry 2013;13:290. 4. Fairburn CG. Cognitive behavior therapy and eating disorders. New York, NY: Guilford Press, 2008. 5. Frisch RE. Fatness and fertility. Sci Am 1988;258:88–95. 6. Krahn DD, Rock C, Dechert RE, Nairn KK, Hasse SA. Changes in resting energy expenditure and body composition in anorexia nervosa patients during refeeding. J Am Diet Assoc 1993;93:434–8. 7. Probst M, Goris M, Vandereycken W, Van Coppenolle H. Body composition in female anorexia nervosa patients. Br J Nutr 1996;76:639–47. 8. Iketani T, Kiriike N, Nagata T, Yamagami S. Altered body fat distribution after recovery of weight in patients with anorexia nervosa. Int J Eat Disord 1999;26:275–82. 9. Grinspoon S, Thomas L, Miller K, Pitts S, Herzog D, Klibanski A. Changes in regional fat redistribution and the effects of estrogen during spontaneous weight gain in women with anorexia nervosa. Am J Clin Nutr 2001;73:865–9. 10. Mayer L, Walsh BT, Pierson RN Jr, Heymsfield SB, Gallagher D, Wang J, Parides MK, Leibel RL, Warren MP, Killory E, et al. Body fat redistribution after weight gain in women with anorexia nervosa. Am J Clin Nutr 2005;81:1286–91. 11. Scalfi L, Polito A, Bianchi L, Marra M, Caldara A, Nicolai E, Contaldo F. Body composition changes in patients with anorexia nervosa after complete weight recovery. Eur J Clin Nutr 2002;56:15–20. 12. Tothill P, James Hannan W. Dual-energy X-ray absorptiometry measurements of fat and lean masses in subjects with eating disorders. Int J Obes Relat Metab Disord 2004;28:912–9. 13. Mayer LE, Klein DA, Black E, Attia E, Shen W, Mao X, Shungu DC, Punyanita M, Gallagher D, Wang J, et al. Adipose tissue distribution after weight restoration and weight maintenance in women with anorexia nervosa. Am J Clin Nutr 2009;90:1132–7. 14. Bodell LP, Mayer LE. Percent body fat is a risk factor for relapse in anorexia nervosa: a replication study. Int J Eat Disord 2011;44:118–23. 15. Mayer LE, Roberto CA, Glasofer DR, Etu SF, Gallagher D, Wang J, Heymsfield SB, Pierson RN Jr, Attia E, Devlin MJ, et al. Does percent body fat predict outcome in anorexia nervosa? Am J Psychiatry 2007; 164:970–2. 16. Mayer L. Body composition and anorexia nervosa: does physiology explain psychology? Am J Clin Nutr 2001;73:851–2. 17. Lee ES, Kim YH, Beck SH, Lee S, Oh SW. Depressive mood and abdominal fat distribution in overweight premenopausal women. Obes Res 2005;13:320–5. 18. Bornstein SR, Schuppenies A, Wong ML, Licinio J. Approaching the shared biology of obesity and depression: the stress axis as the locus of gene-environment interactions. Mol Psychiatry 2006;11:892–902. 19. Albanese CV, Diessel E, Genant HK. Clinical applications of body composition measurements using DXA. J Clin Densitom 2003;6:75– 85. 20. Kang SM, Yoon JW, Ahn HY, Kim SY, Lee KH, Shin H, Choi SH, Park KS, Jang HC, Lim S. Android fat depot is more closely associated with metabolic syndrome than abdominal visceral fat in elderly people. PLoS ONE 2011;6:e27694. 21. Mannucci E, Ricca V, Di Bernardo M, Rotella C. Studio del comportamento alimentare con una intervista strutturata: la Eating Disorder Examination. [Study of eating behavior with a structured interview: The Eating Disorder Examination.] Diabete 1996;8:127–31 (in Italian). 22. Fairburn CG, Cooper Z. The Eating Disorder Examination. 12th ed. In: Fairburn CG, Wilson GT, eds. Binge eating: nature, assessment, and treatment. New York, NY: Guilford Press, 1993:317–60. 23. Dalle Grave R. Multistep cognitive behavioral therapy for eating disorders: theory, practice, and clinical cases. New York, NY: Jason Aronson, 2013. 24. Istituto Nazionale di Ricerca per gli Alimenti e la Nutrizione (INRAN). Linee guida per una sana alimentazione italiana. [Italian Guidelines for
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
Our study did have some limitations. First and foremost, data were obtained in a single inpatient unit by applying one treatment program, which meant that external validation is required. Furthermore, our results need to be interpreted with caution because they may not apply to patients with AN treated in outpatient settings, where a slower rate of weight regain could result in a different regional fat accumulation. However, one outpatient study has already shown a central accumulation of body fat in these subjects, even though weight regain was slower (9). It should also be noted that the less intensive treatment offered in outpatient settings may have a less pronounced effect in terms of mitigating concerns and psychological distress of patients associated with central body fat accumulation. Second, DXA assessment may underestimate lean mass in underweight patients with AN because, as showed by Haas et al (43), patients who have a modest reduction in lean mass can be severely protein depleted. This effect is because starvationinduced changes in body composition may be associated with an increase in total body water, which can, in turn, mask low total body protein when a 3-compartment model, as in DXA, is used. Hence, direct methods may be more reliable means of assessing lean mass in underweight patients. Third, without a long-term follow-up, we were unable to clarify whether lipodystrophy characterized by the preferential accumulation of fat in the central region is a transitory phenomenon, as suggested by a study in 16 patients with AN who a maintained normal body weight 1 y after hospital discharge (13) or whether it persists long-term after weight restoration. Hence, future studies should also take into account the effect of time on body composition by serially assessing patients with AN and healthy subjects because healthy individuals tend to gain weight and abdominal fat as they age. Finally, we did not assess physical activity levels in the control group, which was a feature that could have affected weight and body composition as well as the interpretation of results. In conclusion, our data show that acute weight restoration is associated with a preferential distribution of body fat in central regions, which does not, however, seem to significantly influence eating disorder psychopathology or psychological distress in patients with AN. Nevertheless, future longitudinal studies are needed to clarify whether the preferential accumulation of body fat can influence AN relapse in the long term or the development of metabolic and cardiovascular complications.
777
778
25. 26. 27. 28. 29. 30.
32. 33. 34. 35.
healthy eating.] ed. 2003 (in Italian). Available from: http://www.inran. it/648/linee_guida.html. Stults-Kolehmainen MA, Stanforth PR, Bartholomew JB, Lu T, Abolt CJ, Sinha R. DXA estimates of fat in abdominal, trunk and hip regions varies by ethnicity in men. Nutr Diabetes 2013;3:e64. Mazess RB, Barden HS, Bisek JP, Hanson J. Dual-energy x-ray absorptiometry for total-body and regional bone-mineral and soft-tissue composition. Am J Clin Nutr 1990;51:1106–12. Cooper Z, Cooper PJ, Fairburn CG. The validity of the eating disorder examination and its subscales. Br J Psychiatry 1989;154:807–12. Wilson GT, Smith D. Assessment of bulimia nervosa: an evaluation of the Eating Disorders Examination. Int J Eat Disord 1989;8:173–9. Kendall PC, Marrs-Garcia A, Nath SR, Sheldrick RC. Normative comparisons for the evaluation of clinical significance. J Consult Clin Psychol 1999;67:285–99. De Leo D, Frisoni GB, Rozzini R, Trabucchi M. Italian community norms for the Brief Symptom Inventory in the elderly. Br J Clin Psychol 1993;32:209–13. Derogatis LR, Spencer PM. The Brief Symptom Inventory: administration, scoring, and procedures manual. Baltimore: Clinical Psychometric Research, 1982. Derogatis LR. BSI Brief Symptom Inventory. Administration, scoring, and procedures manual. 4th ed. Minneapolis, MN: National Computer Systems, 1993. Aickin M, Gensler H. Adjusting for multiple testing when reporting research results: the Bonferroni vs Holm methods. Am J Public Health 1996;86:726–8. Holm S. A simple sequentially rejective multiple test procedure. Scand J Stat 1979;6:65–70. Orphanidou CI, McCargar LJ, Birmingham CL, Belzberg AS. Changes in body composition and fat distribution after short-term weight gain in patients with anorexia nervosa. Am J Clin Nutr 1997;65:1034–41.
36. Zamboni M, Armellini F, Turcato E, Todisco P, Gallagher D, Dalle Grave R, Heymsfield S, Bosello O. Body fat distribution before and after weight gain in anorexia nervosa. Int J Obes Relat Metab Disord 1997;21:33–6. 37. El Ghoch M, Alberti M, Milanese C, Battistini NC, Pellegrini M, Capelli C, Calugi S, Dalle Grave R. Comparison between dual-energy X-ray absorptiometry and skinfolds thickness in assessing body fat in anorexia nervosa before and after weight restoration. Clin Nutr 2012; 31:911–6. 38. Carr MC, Brunzell JD. Abdominal obesity and dyslipidemia in the metabolic syndrome: importance of type 2 diabetes and familial combined hyperlipidemia in coronary artery disease risk. J Clin Endocrinol Metab 2004;89:2601–7. 39. Diamanti-Kandarakis E, Bergiele A. The influence of obesity on hyperandrogenism and infertility in the female. Obes Rev 2001;2:231–8. 40. Prioletta A, Muscogiuri G, Sorice GP, Lassandro AP, Mezza T, Policola C, Salomone E, Cipolla C, Della Casa S, Pontecorvi A, et al. In anorexia nervosa, even a small increase in abdominal fat is responsible for the appearance of insulin resistance. Clin Endocrinol (Oxf) 2011; 75:202–6. 41. Bredella MA, Ghomi RH, Thomas BJ, Torriani M, Brick DJ, Gerweck AV, Misra M, Klibanski A, Miller KK. Comparison of DXA and CT in the assessment of body composition in premenopausal women with obesity and anorexia nervosa. Obesity (Silver Spring) 2010;18:2227–33. 42. Ellis KJ. Human body composition: in vivo methods. Physiol Rev 2000;80:649–80. 43. Haas VK, Kohn MR, Clarke SD, Allen JR, Madden S, Mu¨ller MJ, Gaskin KJ. Body composition changes in female adolescents with anorexia nervosa. Am J Clin Nutr 2009;89:1005–10.
Downloaded from ajcn.nutrition.org at MOUNT SAINT VINCENT UNIVERSITY LIBRARY on October 3, 2014
31.
EL GHOCH ET AL
