Inpatient Treatment for Anorexia Nervosa: A Systematic Review of Randomized Controlled Trials

Background. Anorexia nervosa (AN) is a serious psychiatric disease. Choice of acute inpatient care for AN is driven by the severity of symptoms and the level of risk to the patient. Inpatient hospitalization of patients with AN typically includes a behavioral weight gain protocol that is designed to address the core features of the disorder: weight, appetite, and distorted thoughts and behavior. Some add-on treatments may also be included in the inpatient treatment model and may have potential benefits, including faster or greater weight gain; such treatments include psychotherapy, psychoeducation, pharmacological treatment, and nutritional replacement. Objective. The goal of this study was to systematically review randomized clinical trials (RCTs) that have compared the efficacy of different forms of add-on treatment delivered during admission to a 24-hour hospital and to summarize the existing data regarding weight gain associated with such pharmacological, medical, and psychological interventions. Methods. Systematic electronic and manual searches were conducted to identify published RCTs concerning inpatient treatment of AN. Weight gain was used as the main outcome variable. Results. Overall, no significant increase in weight recovery was reported with atypical antipsychotics compared to placebo or therapy as usual. Only one study showed slight benefits in young patients during hospitalization (d = 0.77; 95% confidence interval [CI] –0.09–1.64). No significant effects on weight recovery were found for antidepressants (d = –0.10; 95% CI = –0.63–0.42). In addition, none of the add-on psychotherapy techniques that were evaluated demonstrated superiority compared with control interventions in the inpatient setting. Cyclic enteral nutrition was studied in one RCT in which it demonstrated superiority compared to oral refeeding only (d = 0.97; 95% CI = 0.51–1.47). Other less common treatments such as bright light therapy and lithium carbonate were not found to produce additional significant

Journal of Psychiatric Practice Vol. 21, No. 1

PAULA SUÁREZ-PINILLA, MD CARMEN PEÑA-PÉREZ, DNP BEATRIZ ARBAIZAR-BARRENECHEA, MD BENEDICTO CRESPO-FACORRO, MD, PdD JOSÉ ANDRÉS GÓMEZ DEL BARRIO, MD, PhD JANET TREASURE, MD, PhD JAVIER LLORCA-DÍAZ, MD, PhD

weight improvement compared with placebo. Conclusion. Most add-on treatments during the acute inpatient phase of AN treatment are not effective in increasing weight recovery. Long-term follow-up studies after the acute treatment phase are needed to make evidence-based recommendations. (Journal of Psychiatric Practice 2015;21:49–59) KEY WORDS: anorexia nervosa, hospitalization, weight, body mass index, outcome, add-on interventions

The transfer of care for patients with anorexia nervosa (AN) from internal medicine to psychiatric wards began in the second half of the twentieth century. At that time, the restoration of normal weight was considered to be sufficient treatment.1,2 However, other studies that examined this outcome

SUÁREZ-PINILLA, CRESPO-FACORRO, and DEL BARRIO: University Hospital Marqués de Valdecilla, University of Cantabria, Santander, Spain, Centro Investigación Biomédica en Red Salud Mental, Madrid (CIBERSAM), and Instituto de Investigación Marqués de Valdecilla (IDIVAL), Santander; PEÑAPÉREZ and ARBAIZAR-BARRENECHEA: University Hospital Marqués de Valdecilla, University of Cantabria, Santander; TREASURE: Maudsley Hospital, Denmark Hill, London SE5 8AZ, United Kingdom; LLORCA-DÍAZ: IDIVAL and University of Cantabria, Santander and CIBER EPIDEMIOLOGíA Y SALUD PúBLICA (CIBERESP), Barcelona. Drs. Suárez-Pinilla, Treasure, and Llorca-Díaz contributed equally to this study. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. Please send correspondence to: Paula Suárez-Pinilla, Unidad de Investigación Psiquiatría, Hospital Universitario Marqués de Valdecilla, Avda. Valdecilla, SN. 39008 Santander, Spain. [email protected] This study was performed at the Hospital Marqués de Valdecilla, University of Cantabria, Santander, Spain, with grant support from the Wenceslao López-Albo Fellowship, IFIMAV, Valdecilla Biomedical Research Institute, Santander, Spain. We also wish to acknowledge the contribution of the Eating Disorder Unit (EDU) of the Institute of Psychiatry, London. The authors declare no conflicts of interest. DOI: 10.1097/01.pra.0000460621.95181.e2

January 2015

Copyright © Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

49

INPATIENT TREATMENT FOR ANOREXIA NERVOSA

in the long-term found that relapse was common.3 A 2013 study that examined the predictive value of a number of outcomes found that achieving an ideal body weight at the end of the treatment is considered one of the best predictors of recovery in patients with AN.4 However, although weight restoration is the first objective of therapy, it is not a sufficient condition for recovery.4–6 Given the medical risks associated with the symptoms of the illness and the risk of a fatal outcome, inpatient treatment is widely used for patients with AN. Hospitalization for AN typically involves a behavioral weight gain protocol, and inpatient treatment is often considered the most reliable way to achieve weight gain in this population, given the ability to provide a controlled environment.7 The model of inpatient care for AN has changed in recent years. Over the 15-year period between 1984 and 1998, the average length of admission decreased from 149.5 days to 23.7 days, and there was a decrease from 19 kg/m2 to 17 kg/m2 in mean body mass index (BMI) at discharge.8 This is important because the risk of relapse and readmission increases as the likelihood of achieving ideal body weight at discharge decreases.9 The decrease in BMI and length of stay in recent years occurred in the context of increased emergence of day programs and partial hospitalization. Consequently, patients typically start their weight restoration in inpatient programs, and they are then transferred to day hospitalization after discharge. Day patient programs have been demonstrated to be effective in maintaining goals achieved during hospitalization and producing further improvement, making these programs an alternative for those patients who are medically stable.10 The form and content of inpatient care for AN have also changed over time. Initially, nursing care followed strict behavioral principles to manage weight gain. However, a more lenient program was subsequently found to be more acceptable to patients and staff, with no significant difference in rate of weight gain.11 Meal support and re-feeding are key components of inpatient care, but there is uncertainty about how these strategies are best delivered and both nurses and patients find these methods difficult.12 As noted above, inpatient hospitalization of patients with eating disorders typically includes a behavioral weight gain protocol that is designed to

50

January 2015

address the core features of AN: weight, appetite, and distorted thoughts and behavior. Some add-on treatments may also be included in the inpatient treatment model and may have potential benefits, including faster or greater weight gain. These options include, among others, psychotherapy, psycho-education, pharmacological treatment, and nutritional replacement. However, although a variety of add-on treatments for AN have been tried, few randomized controlled trials (RCTs) have been carried out and few favorable results have emerged. The goal of this study was to systematically review RCTs that have compared the efficacy of different forms of add-on treatment delivered during admission to a 24-hour hospital and to provide a summary of the existing data regarding weight gain associated with such pharmacological, medical, and psychological interventions.

METHODS Comprehensive Literature Search We searched electronic databases, including the Cochrane Database of Systematic Reviews, PsycINFO, Medline, Embase, Web of Science, Science Citation Index, and the Cochrane Central Register of Controlled Trials. Manual searches of journals that publish research on the treatment of eating disorders and of reference lists from previous reviews were also carried out. Eligibility The literature review was limited to articles in English and Spanish published during the last 40 years through September 2013. The main search terms were inpatient treatment AND anorexia nervosa. Linked search terms were therapy, treatment, tube feeding, nutrition, pharmacotherapy, fluoxetine, antidepressant, olanzapine, quetiapine, neuroleptic, psychotherapy, family therapy, cognitive therapy, behavior therapy, self-help, group therapy AND prognosis, outcome, weight gain, BMI. The abstracts (or executive summaries) from identified articles were reviewed independently by two of the authors (PS-P and JT) to determine eligibility for inclusion in the review. Disagreements over articles, quality, and inclusion criteria were addressed through discussions with the co-authors of this article.

Journal of Psychiatric Practice Vol. 21, No. 1

Copyright © Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

INPATIENT TREATMENT FOR ANOREXIA NERVOSA

Study Quality and Risk of Bias The methodological quality of the studies was independently assessed by three of the authors (PS-P, CPP, and BA-B). The risk of bias was examined by using the Cochrane risk bias tool13 for randomized trials. Heterogeneity between articles was assessed using the DerSimonian and Laird test.14 Any discrepancies were evaluated and resolved by a fourth reviewer (JL-D). In addition, Egger’s regression15 and Begg’s test16 were used to assess for publication bias. Data Extraction

item in the studies included in our analysis. Weight gain was mainly reported in terms of BMI,20 percentage of ideal BMI, or weight increase over time. The primary outcomes are summarized in Table 1. Secondary outcome variables varied among the studies and are described in the results. Statistical Analyses Analyses were performed using Microsoft Excel and EPIDAT 4.0 software. After a thorough evaluation of the studies, the individual effect size of weight at discharge was calculated as the standardized mean difference (SMD), when possible, comparing active treatment with placebo or treatment as usual (TAU), as the reference. The global effect size, Cohen’s d, was assessed combining the SMDs and confidence intervals from each study, using the inverse variance method. Cohen suggested that d = 0.2 be considered a “small” effect size, 0.5 a “medium” effect size, and 0.8 a “large” effect size. To consider statistical significance, the 95% confidence interval should not contain zero. The cut off point for statistical significance was a P value of 0.05. Because of the variability among the different studies, a random effect model was chosen to interpret the results. The majority of studies did not provide sufficient information about potential confounders such as type of AN (restricting or binge-eating/purging type), length of illness, length of admission, or other sources of heterogeneity; therefore, subgroup analysis for these variables of interest could not be performed.

Information about the design and participants in each study was extracted based on the Preferred Reporting Items for Systematic Reviews and MetaAnalyses (PRISMA).17 Data were extracted from the source documents independently by four of the authors: three physicians (PS-P, JT, and BC-F) and one epidemiologist (JL-D). RCTs that met the following inclusion criteria were included: 1. Clinical diagnosis of AN, restricting or binge-eating/purging type, or eating disorder not otherwise specified (EDNOS), according to the criteria in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IVTR)18 or the International Statistical Classification of Diseases, Tenth Edition (ICD-10)19 2. Study performed in a 24-hour hospital 3. Data extracted in the study included the following information: a) randomization with assessment of differences in characteristics between randomized groups at baseline, b) ethical approval (eg, approval by an institutional review board), c) description of the intervention, d) complete description of the form of analysis (eg, completion/intention to treat), e) side effects, f) primary outcome measure of weight, BMI, or weight gain, and g) secondary outcome measures, when available: eating attitude and behavior, cost effectiveness, and general psychopathology. Studies were excluded if the subjects met DSM-IV-TR criteria for bulimia nervosa or if they were conducted completely in an outpatient or day hospital setting.

Electronic database searches yielded 931 references. Figure 1 describes the flow diagram of the study selection process. After exclusion, 18 RCTs were included in the global synthesis. Table 1 presents sociodemographic and descriptive data from the studies included in the analysis and Table 2 presents data on outcome effect sizes. The gender of the participants in all of the studies was more than 95% female.

Outcome Measures

Atypical Antipsychotics

We chose weight gain as the primary outcome criterion because this was the most commonly reported

Study setting. We found 4 relevant studies concerning atypical antipsychotics, 2 of which were conduct-

Journal of Psychiatric Practice Vol. 21, No. 1

RESULTS Study Selection

January 2015

Copyright © Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

51

52

January 2015

Copyright © Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

AN-R-BP, EDNOS

AN-R-BP

AN-R-BP

AN-R-BP

AN-R, EDNOS

AN-R-BP

AN-R-BP

AN-R-BP

Crisp et al 198727

Attia et al 199828

Eckert et al 197929

Geist et al 200030

Pillay & Crisp 198131

Davies et al 201232

Rigaud et al 200733

Birmingham et al 199434 AN-R-BP

AN-R

Holtkamp et al 200526

Kells et al 201235

Janas-Kozik et al 201136 AN-R

16 / 13

8/8

46 / 33

11 / 12

12 / 13

40 / 41

31 / 22

8/8

19 / 13

14 / 16

18 / 22

10 / 10

10 / 11

21.5 (8) / 22.5 (2)

20.6 (1.8) / 18.6 (2.6)

17.8 (1.3) / 17.0 (1.3)

15.9 (2.5)

20.6 (3.8) / 24.8 (6.1)

22.5 (4.5) / 24.2 (3.8)

26.1 (7.6) / 26.0 (7.9)

23.6 (8.2) / 23.8 (7.8)

14.3 (1.5) / 14.9 (1.7)

Nq

29.1 (7.2) / 23.4 (6.4)

20.9 / 21.3

14.8 (1.2) / 13.9 (1.3)

20.56 (5.1)

16.2 (2.5) / 15.8 (2.3)

14.4 (2.2) / 18.0 (2.1)

23.8 (9.4) / 21.0 (3.3)

13.3 (2.2) / 14.42 (1.8)

Age, years mean (SD)

23.5m (4.8) / 39.5m (11.4)

Nq

Nq

Nq

3.6y (2.0) / 2.6y (2.7)

4.5y (1.9) / 3.2y (2.7)

10.6y (7) / 9.9y (7.6)

> 5y

Nq

Nq

8y (5.8)

5y / 3.7y

12.0m (7.0) / 10.0m (7.0)

2.9y (2.3)

Nq

> 4y

65.4m (96.2) / 30.3m (37.3)

10.6m (6.9) / 11.6m (8.5)

Duration of illness mean (SD)

3–4

8

10

4

6–7

5

7

10

15

4

11

10

12

2

Duration, weeks

6

Cisapride / placebo

8

CBT+lithium / CBT+placebo 4

CBT+BLT / CBT

Supervision / no supervision 1

Zinc / placebo

CEN + meals / meals

CREST / TAU

Social skills / TAU

Family therapy / psychoeducation

Behavior therapy / TAU

Fluoxetine / placebo

Clomipramine / placebo

SSRI / control

Amitriptyline / placebo

Risperidone / placebo

Olanzapine / placebo

Quetiapine / TAU

Olanzapine / control

Treatment

Abbreviations: AN: anorexia nervosa;BLT: bright light therapy; BP: binge-eating/purging type; CBT: cognitive-behavioral therapy; CEN: cyclic enteral nutrition; CG: control group; CREST: cognitive remediation and emotional skills training; EDNOS: eating disorder not otherwise specified; EG: experimental group; m: months; Nq: not quantified; R: restricting type; SD: standard deviation; SSRI: selective serotonin reuptake inhibitor; TAU: therapy as usual; y: years.

AN-R-BP

12 / 12

AN-R

Halmi et al 198625

Szmukler et al 199538

13 / 39

AN-R-BP

Hagman et al 201124

AN-R-BP

16 / 19

AN-R

Kafantaris et al 201123

Gross et al 198137

41 / 40

AN-R-BP

Court et al 201022

11 / 11

AN-R-BP, EDNOS

Norris et al 201121

N

Diagnosis

Study

Experimental group / Control group

Table 1. Sociodemographic and descriptive data from the studies included in the systematic review

INPATIENT TREATMENT FOR ANOREXIA NERVOSA

Journal of Psychiatric Practice Vol. 21, No. 1

INPATIENT TREATMENT FOR ANOREXIA NERVOSA

Figure 1. Flow diagram of selection process 931 references retrieved from bibliographic search 215 references were duplicated and/or unlikely to be relevant based on title and abstract 716 full-text articles retrieved for detailed evaluation 698 articles excluded: 201 reviews or meta-analyses 51 letters to the editor or comments 116 case reports 155 not randomized controlled trials 175 outpatient settings 18 articles for final inclusion in the review: 4 antipsychotic medications 4 antidepressant medications 4 psychotherapy interventions 3 nutrition procedures 3 other treatments

ed completely in an inpatient setting,21,22 1 of which involved a mixed sample that was mostly inpatients (15 inpatients, 5 outpatients),23 and 1 in which all of the subjects were initially admitted to an inpatient or intensive day-hospital treatment level of care.24 Participants. One study included patients with AN restricting type only,23 whereas the other three antipsychotic studies included both AN types.21,22,24 One study also included a group of patients with EDNOS on the anorexic spectrum.21 Two studies21,23 were conducted in patients under 16 years of age, with a mean duration of illness of less than 1 year in one study21 but longer than 4 years in the other.23 The other 2 studies were carried out in patients with a mean age of 16 years or higher and a mean duration of illness longer than 3 years.22,24 Baseline differences. In one study, randomization was stratified by inpatient versus outpatient status at study entry to assess for possible differences between these subgroups.23 In the study by Norris et al., the investigators noted that patients treated with olanzapine presented with greater acuity and

Journal of Psychiatric Practice Vol. 21, No. 1

more complex psychopathology than those patients not treated with olanzapine, making comparisons regarding efficacy of the drug impossible.21 Intervention. Two studies, one of which involved olanzapine23 and one risperidone,24 compared the addition of an atypical antipsychotic medication versus placebo to the treatment regimen. One study compared patients treated with olanzapine with a matched sample that did not receive medication treatment,21 and one compared patients treated with quetiapine with those receiving TAU.22 All of the studies employed a low to normal dose of the antipsychotic drug: 5–10 mg of olanzapine, 100–400 mg of quetiapine, and 0.5–4.0 mg of risperidone daily. All of the inpatient programs provided medical care, nutrition management, individual meal plans for weight restoration, and psychological treatment consisting of individual cognitive-behavioral therapy (CBT), group, family, and multifamily group therapies, and other psychological interventions.21–24 Primary outcome. Changes in BMIs from patient admission to discharge were calculated to assess the primary outcome for our review. The global SMD in changes in BMIs between the groups treated with antipsychotic medication (olanzapine or quetiapine) and the control groups was not significant (d = 0.32; 95% confidence interval (CI), –0.18–0.81).21–23 The addition of risperidone to the treatment regimen did not demonstrate efficacy in improving weight restoration in adolescents.24 The addition of olanzapine to the treatment regimen produced a slight improvement in BMI in adolescent subjects (d = 0.77; 95%CI, –0.09–1.64), but this difference was not significant.21 Secondary outcomes. In adolescents, the length of stay was significantly longer in the group in which olanzapine was added to the treatment regimen (t = 2.68; P = 0.023).21 In the studies conducted in patients with a mean age over 16 years, the addition of risperidone and quetiapine to the treatment regimen was associated with reduced eating disorder psychopathology.22,24 especially during the early weeks of treatment. Side effects and adverse events. Dyslipidemia and sedation were observed in 56% of the young patients with AN who were treated with olanzapine in the Norris et al study,21 and a trend toward increased fasting glucose and insulin levels was found in the adolescents treated with olanzapine in the Kafantaris et al study.23 Risperidone was associated

January 2015

Copyright © Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

53

54

January 2015

Copyright © Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

Weight gain (kg)

Weight (kg)

BMI

—

37.7 (1.1) / 33.9 (1.9)

14.7 (1.2) / 15.9 (2.4)

— — —

74.9 (9.2) / 77.2(11.1) — —

Nq

— 14.3 (1.2) / 14.8 (1.4) — 72.5 (5.3) / 71.8 (5.0)

15.6 (1.6) / 15.5 (1.5) 16.9 (1.7) / 16.3 (1.8) 16.9 (0.6) / 16.0 (1.5) 15.9 (1.9) / 16.1 (1.3)

Primary outcome Mean (SD) baseline

0.38 (–0.34–1.11) 0.13 (–0.58–0.84) –1.15 (–2.21 to –0.09) –0.15 (–0.75–0.45)

0.77 (–0.09–1.64) –0.04 (–0.89–0.81) 0.23 (–0.65–1.10) Nq

5.7 (0.6) / 5.1 (0.5)

43.0 (1.1) / 37.8 (1.9)

10% BMI increase from baseline in both groups

194 (14) / 126 (99) 3.6 (2) / 2.6 (7) 0.35 (0.23) / 0.33 (0.29)

—

—

Nq

0.97 (0.51–1.43) 0.19 (–0.48–0.85) 0.07 (–0.58–0.70)

–1.09 (–1.88 to –0.31)

0.64 (–0.37–1.64)

Nq

0.44 (–0.16–1.05)

–0.77 (–2.16–0.62)

Nq

–0.10 (–0.63–0.42)

Nq

0.32 (–0.18–0.81)

Individual effect size Global effect size d (CI) d (CI)

No significant difference Nq in weight gain between groups 93.3 (7.3) / 96.3 (8.2) –0.05 (–0.84–0.73) 34.5 (15.5) / 37.1 (18.6) –0.15 (–0.97–0.67) 1.3 (0.2) / 1.8 (0.3) –2.03 (–2.58 to –1.48)

2.1(1.5) / 1.5 (1.2) 17.0 (0.9) / 17.3 (0.8) 0.17 (0.04) / 0.21 (0.03) 87.0 (5.5) / 87.4 (4.7)

16.6 (1.8) / 18.3 (0.9) 18.6 (1.8) / 18.1 (1.8) 18.1 (2.0) / 17.7 (1.8) No significant difference in weight restoration between groups

Primary outcome Mean (SD) end

Abbreviations: BMI: body mass index; d: Cohen´s d; CI: confidence interval; Nq: not quantified; SD: standard deviation *Treatment efficiency equals the reciprocal of days to target weight multiplied by 90.

Szmukler et al 199538

Cisapride

Gross et al 198137

Lithium carbonate

Jana-Kozik et al 201136

Bright light therapy

Rigaud et al 200733 Birmingham et al 199434 Kells et al 201235 Weight gain (gm / day) Total weight gain (kg) Weight gain (kg / day)

% ideal body weight % increased weight % increased BMI

Geist et al 200030 Pillay & Crisp 198131 Davies et al 201232

Nutrition

Averageweight gain

Treatment efficiency* BMI Rate of weight gain % ideal body weight

BMI BMI BMI BMI

Primary outcome

Eckert et al 197929

Psychotherapy

Halmi et al 198625 Holtkamp et al 200526 Crisp et al 198727 Attia et al 199828

Antidepressants

Norris et al 201121 Court et al 201022 Kafantaris et al. 201123 Hagman et al 201124

Antipsychotics

Study

Experimental group / Control group

Table 2. Outcomes in the studies included in the systematic review

INPATIENT TREATMENT FOR ANOREXIA NERVOSA

Journal of Psychiatric Practice Vol. 21, No. 1

INPATIENT TREATMENT FOR ANOREXIA NERVOSA

with a non-significant increase in prolactin levels at the end of the study by Hagman et al.24 Minimal side effects were reported with a low dose of quetiapine.22 Antidepressant Therapy Study setting. We identified 4 studies carried out in an inpatient setting that compared antidepressant medications to placebo or TAU.25–28 Participants. Three studies included a mixture of both AN types,26–28 one of which also included a minority of patients with EDNOS on the anorexic spectrum.26 One study involved only patients with AN restricting type.25 The length of illness ranged between 2.9 and 8 years in the adult samples25,27,28 and less than one year in the adolescent sample.26 Intervention. Three of the RCTs involved comparisons of active drugs—fluoxetine,28 amitriptyline,25 and clomipramine27—with placebo and one compared patients treated with selective serotonin reuptake inhibitors (SSRIs) with an untreated matched sample.26 Doses of the antidepressants were 160 mg amitriptyline, 50 mg clomipramine, 20–56 mg fluoxetine, 120 mg fluvoxamine, and 100 mg sertraline, daily. All of the patients in these studies were also being treated with behavioral weight gain protocols.25–28 Participants attended a variety of group and individual therapy sessions and received weight and nutritional management, dietician therapy, and family counselling. In the case of insufficient weight gain, reinforcement strategies or nasogastric tubefeeding were employed to try to achieve the target weight.26 Baseline differences. In the study examining the adjunctive use of fluoxetine, the patients who received fluoxetine were slightly older than the control group. In that study only, randomization was stratified by AN type and by the presence or absence of major depression.28 In the study examining the use of SSRIs in adolescents with AN, the group who were randomized to receive an SSRI had higher levels of core eating disorder psychopathology and depressive symptoms than the control group at the time that the medication was started.26 Primary outcome. No benefits in terms of weight gain were found in these four studies (d = –0.10; 95% CI, –0.63–0.42). Secondary outcomes. Patients treated with clomipramine were found to have more stable eating habits.27

Journal of Psychiatric Practice Vol. 21, No. 1

Side effects and adverse events. The most common physical complaints with amitriptyline were drowsiness, excitement, confused state, increased motor activity, dry mouth, and constipation.25 Fluoxetine was in general well tolerated, but one of the patients reported insomnia, and another had blurred vision.28 Psychotherapy Study setting. Four studies compared different forms of psychological therapy with TAU or psychoeducation,29–32 all of which were conducted in an inpatient setting. Participants. Three of the studies were carried out in patients with both AN types; one involved patients with restricting AN or restricting EDNOS.30 In one RCT, the patients were under 18 years of age,30 one study did not specify the age of the patients,29 and the other two studies were conducted in an adult treatment setting, with a mean duration of illness of more than 5 years.31,32 Intervention. Three of these RCTs compared TAU with a form of psychotherapy: behavioral therapy,29 social skills training,31 and cognitive remediation and emotion skills training (CREST).32 The fourth study compared family therapy versus family group psychoeducation.30 TAU varied across the studies. In two studies,30,31 the usual treatment consisted of a behavioral program with initial bed rest, individual and family psychotherapy, and milieu therapy. In one study, TAU included nutrition management, dietetic input, occupational therapy, family work, individual work, and group work.32 In one study, all subjects received the same personal therapy program, with or without behavior modification.29 Baseline differences. In the study investigating social skills training, the patients who received TAU had slightly lower baseline scores on obsession and depression scales than those who received social skills training.31 Primary outcome. The global SMD in body weight gain between the control groups and the groups who received the adjunctive psychotherapy treatment (social skills training, family therapy, and CREST) was not significant (d = -0.77; 95% CI, –2.16–0.62).30–32 However, both experimental and control groups showed a significant effect for restoration of body weight.30–32 Behavior therapy was associated with a significant difference in weight gain

January 2015

Copyright © Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

55

INPATIENT TREATMENT FOR ANOREXIA NERVOSA

compared with TAU only in the subgroup of patients who had received no prior outpatient treatment.29 Secondary outcomes. Social skills training was associated with improved ability to stay in the overall treatment during the inpatient period.31 The TAU and CREST groups both showed significantly improved cognitive flexibility; the CREST group also showed an improvement in central coherence.32 Side effects and adverse events. No relevant side effects or adverse events were reported in these studies. Nutrition Study setting. Three studies assessed nutrition treatment in an inpatient setting.33–35 Participants. One study included only patients who were younger than 18 years of age with the restrictive AN.35 The two other studies included patients with both binge-eating/purging and restricting types of AN who were being treated in an adult setting.33,34 The average duration of illness in the studies in adult settings was longer than 3 years. Intervention. One RCT compared cyclic enteral nutrition (CEN) added to oral intake versus oral intake alone. The daily flow rate was 1.5−2.5 ml/min with CEN, using iso-nitrogen iso-energetic mixtures. The quantity was adapted to oral intake.33 One study compared supplementation with 100 mg of zinc daily with placebo,34 and the third study compared unsupervised meals with meals supervised by nurses.35 Inpatient treatment consisted of dietary interventions, behavioral therapy, and psychotherapy.33–35 In the study of zinc supplementation, changes in the intensity of the care provided were structured following a program with different stages based on the patient’s clinical needs.34 Baseline differences. No significant baseline differences were reported. Primary outcome. There were no significant differences in the global SMD with regard to weight gain among the groups in the 3 studies (d = 0.44; 95% CI, –0.16–1.05).33–35 However, the addition of CEN to oral intake significantly improved weight in one study (d = 0.97; 95% CI, 0.51–1.43).33 Zinc supplementation of anorexia nervosa was also reported to increase patients’ weight gain (P = 0.03) in one study.34 Secondary outcomes. The relapse-free period after discharge was longer in the CEN group than in the control group.33 There was an improvement in

56

January 2015

overnight heart rate trends for patients who received at least one supervised meal each day compared with those who received no supervised meals.35 Side effects and adverse events. No relevant side effects or adverse events were reported in these studies. Other Treatments: Bright Light Therapy, Pro-Kinetic Agents, and Lithium Carbonate Study setting. We identified three RCTs for other types of treatments conducted in an inpatient setting.36–38 Participants. Two studies involved patients over 18 years of age with both AN types.37,38 The third study involved patients between 15 and 20 years of age with AN restricting type and depressive symptoms.36 The duration of illness was only specified in one of the three studies, where the average length of illness was greater than 2 years.38 Intervention. One RCT was completed for each of the following types of treatments during inpatient admission: bright light therapy (BLT) combined with CBT compared with CBT,36 the addition of lithium carbonate versus placebo to CBT, 37 and the addition of a prokinetic drug (cisapride) versus placebo to the treatment regimen.38 The dose of BLT was 10,000 lux for 30 minutes daily. The acceptable plasma level of lithium in those patients treated with lithium carbonate was 1.0 ± 0.1 mEq/l, and 10 mg of cisapride was administered three times daily. The independent treatment in each study was a behavior modification program.36–38 Baseline differences. Baseline caloric intake was higher in the group that received lithium.37 Primary outcome. The SMDs between subjects treated with cisapride (d = –1.09; 95% CI, –1.88 to –0.31) and lithium (d = 0.64; 95% CI, –0.37–1.64], and the control groups were not significant in terms of weight. Patients treated with BLT plus CBT experienced an earlier start of weight gain compared to those who received CBT alone, but there were no differences in BMI gain after 6 weeks.36 Secondary outcomes. The combination of BLT and CBT compared to CBT alone led to a significant reduction in depressive symptoms.36 Lithium appeared to provide some improvement in insight into illness.37 Cisapride was associated with changes in subjective measures, such as greater improvement in hunger sensations.38

Journal of Psychiatric Practice Vol. 21, No. 1

Copyright © Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

INPATIENT TREATMENT FOR ANOREXIA NERVOSA

Side effects and adverse events. No relevant side effects or adverse events were reported in these three studies. Heterogeneity and Publication Bias The DerSimonian and Laird (Q) test for subgroups14 rejected heterogeneity between included articles (all 2  5.89; all P  0.12) No publication bias was detected from either Begg’s test16 (all Z  –0.34; all P  0.73) or Egger’s test15 (all t  –1.23; all P  0.34).

DISCUSSION This review of the therapeutic effectiveness of addon treatments for AN when provided in an inpatient setting revealed that few treatments are adequately effective in normalizing weight in acute phases of AN. A variety of controversial treatments have been studied and, in most cases, no differences in weight recovery were found between the placebo and experimental groups. However, the inability to demonstrate the efficacy of a treatment that is added on to a behavioral weight protocol that is usually effective does not imply that the experimental intervention, if used in the absence of the baseline treatment, would not be effective. Overall, the studies involving atypical antipsychotics did not support a significant weight improvement during hospitalization in patients receiving those agents compared to the control groups. Our general findings are consistent with two recent meta-analyses of studies carried out in inpatient and outpatient settings39,40 that failed to demonstrate antipsychotic efficacy for BMI recovery in patients with AN. Olanzapine may be slightly beneficial in facilitating weight restoration in adolescents, but it should not be prescribed as a primary treatment due to the potential for metabolic side effects.41 The use of antipsychotics, including second generation agents, is commonly associated with weight gain.42 However, recent studies have demonstrated that this feature depends on several factors, such as environmental and eating- and lifestyle-related factors.43 On the other hand, antidepressant drugs did not produce weight gain benefits during hospitalization compared to placebo. The period of hospitalization in the studies we examined concerning antidepressant therapy ranged from 4 to 15 weeks. Therefore, it

Journal of Psychiatric Practice Vol. 21, No. 1

seems unlikely that the lack of efficacy of antidepressant treatment that we found in our review was due to a short length of treatment. Moreover, few studies have examined the efficacy of antidepressant medications in eating disorders, and most of those studies have demonstrated efficacy only for the treatment of comorbid disorders such as depression and obsessive compulsive disorder.44 The lack of effectiveness in weight recovery of additional psychotherapy compared with TAU may be explained, in part, by the fact that only a limited number of interventions can be carried out during hospitalization. Hence, family therapy usually has benefits after 6–12 months.30 Moreover, it has been postulated that mere hospitalization of a patient with AN may frequently have a positive impact on weight gain.25 Intensive nutritional rehabilitation, using CEN, is one of the most effective add-on interventions in terms of weight restoration. A recent study demonstrated that a higher caloric intake in hospitalized adolescents with AN is also associated with reduced length of stay and no increased rate of re-feeding syndrome.45 However, although CEN was demonstrated to be a useful support for hospitalized patients with AN, several authors prefer to employ oral re-feeding in most cases and to reserve tube feeding for very malnourished patients because of the psychological side effects and the difficulty of maintaining long-term weight after discharge.46,47 One study also found that zinc supplementation was a viable option to assist in weight restoration in AN, given its low cost and low rate of side effects.34 Non-significant weight benefits were found with the use of alternative treatments such as BLT, cisapride, and lithium carbonate. Even lithium carbonate was associated with a non-significant weight improvement; this result may be biased because patients in the experimental group had higher caloric intake at baseline.37 Cisapride is not used in AN since it was voluntarily withdrawn from the market in the United States and most European countries because it has risks of serious cardiac effects.48 Patients with AN frequently exhibit symptoms of mood disturbance and obsessive-compulsive features that impair psychological functioning.49 Given the frequency with which general psychopathology is comorbid with AN symptoms, some treatments, even though they do not produce benefits in weight gain, may improve other psychopathological outcomes.

January 2015

Copyright © Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

57

INPATIENT TREATMENT FOR ANOREXIA NERVOSA

Hence, BLT reduced depressive symptoms, supporting theories about the effect of sunlight on serotoninergic activity.50 Risperidone and quetiapine also led to a decrease in eating disorder psychopathology.22,24 However, even though antidepressant medications have demonstrated effectiveness in the treatment of major depression, SSRIs have shown limited efficacy in patients with AN,51 and secondary effects in adolescents remain unclear.26 To our knowledge, our study is the largest systematic effort to quantitatively summarize data about hospital treatments for patients with AN. However, a number of limitations of this work should be noted. First, a variety of different types of studies were included in the analyses, which could limit the statistical power, even though we tested for heterogeneity. Therefore, subgroup analyses depending on the form of intervention were done. Second, we used weight restoration as the primary outcome for measuring effectiveness because it was a common outcome in all of the RCTs we analyzed. However, BMI and weight usually offer concrete and easily measured markers and are the primary goal of hospitalization.52 Finally, it is worth noting that our results are based on patients who required hospital admission, who may not be representative of individuals with a less severe form of AN. In conclusion, inpatient treatments remain understudied, and, in many respects, choosing the most appropriate and effective interventions for AN is still a challenge today. Our findings indicate that further efforts to develop new treatment strategies for the acute phase of AN are needed, with long-term followup required to evaluate for objective improvements.

7.

8.

9.

10.

11.

12.

13.

14. 15.

16.

17.

18.

19.

20.

References 1.

2.

3.

4.

5. 6.

58

Treasure P, Hoek H, Palmer B. Eating disorders across Europe. European Eating Disorder Review. 2001;9: 215–216. Russell GF, Szmukler GI, Dare C, et al. An evaluation of family therapy in anorexia nervosa and bulimia nervosa. Arch Gen Psychiatry. 1987;44:1047–1056. Morgan HG, Purgold J, Welbourne J. Management and outcome in anorexia nervosa. A standardized prognostic study. Br J Psychiatry. 1983,143:282–287. Lock J, Agras WS, Le Grange D, et al. Do end of treatment assessments predict outcome at follow-up in eating disorders? Int J Eat Disord. 2013;46:771–778. Hsu LKG. Eating Disorders. New York: Guilford; 1990. Rausch Herscovici C. Lunch session, weight gain and their interaction with the psychopathology of anorexia nervosa

January 2015

21.

22.

23.

24.

in adolescents. Vertex. 2006;17:7–15. Bodell LP, Keel PK. Current treatment for anorexia nervosa: Efficacy, safety and adherence. Psychol Res Behav Manag. 2010;3:91–108. Wiseman CV, Sunday SR, Klapper F, et al. Changing patterns of hospitalization in eating disorder patients. Int J Eat Disord. 2001;30:69–74. Baran SA, Weltzin TE, Kaye W. Low discharge weight and outcome in anorexia nervosa. Am J Psychiatry. 1995;152: 1070–1072. Willinge AC, Touyz SW, Thornton C. An evaluation of the effectiveness and short-term stability of an innovative Australian day patient programme for eating disorders. Eur Eat Disord Rev. 2010;18:220–233. Touyz SW, Beumont PJ, Glaun D, et al. A comparison of lenient and strict operant conditioning programmes in refeeding patients with anorexia nervosa. Br J Psychiatry. 1984;144:517–520. Long S, Wallis D, Leung N, et al. All eyes are on you: Anorexia nervosa patient perspectives of in-patient mealtimes. J Health Psychol. 2012;17:419–428. Higgins JPT, Altman DG, Gøtzsche PC, et al. The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. BMJ. 2011;343:d5928. DerSimonian R, Laird N. Meta-analysis in clinical trial. Control Clin Trials. 1986; 7:177–188. Egger M, Smith GD, Schneider M, et al. Bias in metaanalysis detected by a simple, graphical test. BMJ. 1997;315:629–634. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994;50:1088. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. BMJ. 2009;339:b2535. American Psychiatric Association (APA). Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR). Washington, DC: APA; 2000. World Health Organization. International Statistical Classification of Diseases and Health Related Problems (ICD-10). Geneva: World Health Organization; 2004. American Psychiatric Association Work Group on Eating Disorders. Practice guideline for the treatment of patients with eating disorders (revision). Am J Psychiatry. 2000;157(1 Suppl):1–39. Norris ML, Spettigue W, Buchholz A, et al. Olanzapine use for the adjunctive treatment of adolescents with anorexia nervosa. J Child Adolesc Psychopharmacol. 2011;21: 213–220. Court A, Mulder C, Kerr M, et al. Investigating the effectiveness, safety and tolerability of quetiapine in the treatment of anorexia nervosa in young people: A pilot study. J Psychiatr Res. 2010;44:1027–1034. Kafantaris V, Leigh E, Hertz S, et al. A placebo-controlled pilot study of adjunctive olanzapine for adolescents with anorexia nervosa. J Child Adolesc Psychopharmacol. 2011;21:207–212. Hagman J, Gralla J, Sigel E, et al. A double-blind, placebocontrolled study of risperidone for the treatment of adolescents and young adults with anorexia nervosa: A pilot

Journal of Psychiatric Practice Vol. 21, No. 1

Copyright © Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

INPATIENT TREATMENT FOR ANOREXIA NERVOSA

25.

26.

27.

28.

29. 30.

31.

32.

33.

34.

35.

36.

37.

38.

study. J Am Acad Child Adolesc Psychiatry. 2011;50: 915–924. Halmi KA, Eckert E, LaDu TJ, et al. Anorexia nervosa. Treatment efficacy of cyproheptadine and amitriptyline. Arch Gen Psychiatry. 1986;43:177–181. Holtkamp K, Konrad K, Kaiser N, et al. A retrospective study of SSRI treatment in adolescent anorexia nervosa: Insufficient evidence for efficacy. J Psychiatr Res. 2005;39:303–310. Crisp AH, Lacey JH, Crutchfield M. Clomipramine and ‘drive’ in people with anorexia nervosa: An in-patient study. Br J Psychiatry. 1987;150:355–358. Attia E, Haiman C, Walsh BT, et al. Does fluoxetine augment the inpatient treatment of anorexia nervosa? Am J Psychiatry. 1998;155:548–551. Eckert ED, Goldberg SC, Halmi KA, et al. Behaviour therapy in anorexia nervosa. Br J Psychiatry. 1979;134:55–59. Geist R, Heinmaa M, Stephens D, et al. Comparison of family therapy and family group psychoeducation in adolescents with anorexia nervosa. Can J Psychiatry. 2000;45: 173–178. Pillay M, Crisp AH. The impact of social skills training within an established in-patient treatment programme for anorexia nervosa. Br J Psychiatry. 1981;139:533–539. Davies H., Fox J., Naumann U. et al. Cognitive remediation and emotion skills training for anorexia nervosa: An observational study using neuropsychological outcomes. Eur Eat Disord Rev. 2012;20:211–217. Rigaud D, Brondel L, Poupard AT, et al. 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;26:421–429. Birmingham CL, Goldner EM, Bakan R. Controlled trial of zinc supplementation in anorexia nervosa. Int J Eat Disord. 1994;15:251–255. Kells M, Davidson K, Hitchko L, et al. Examining supervised meals in patients with restrictive eating disorders. Appl Nurs Res. 2012;26:76–79. Janas-Kozik M, Krzystanek M, Stachowicz M, et al. Bright light treatment of depressive symptoms in patients with restrictive type of anorexia nervosa. J Affect Disord. 2011; 130:462–465 Gross HA, Ebert MH, Faden VB, et al. A double-blind controlled trial of lithium carbonate [in] primary anorexia nervosa. J Clin Psychopharmacol. 1981,1:376–381. Szmukler GI, Young GP, Miller G, et al. A controlled trial of cisapride in anorexia nervosa. Int J Eat Disord. 1995;

Journal of Psychiatric Practice Vol. 21, No. 1

17:347–357. 39. Lebow J, Sim LA, Erwin PJ, et al. The effect of atypical antipsychotic medications in individuals with anorexia nervosa: A systematic review and meta-analysis. Int J Eat Disord. 2013,46:332–339 40. Kishi T., Kafantaris V, Sunday S, et al. Are antipsychotics effective for the treatment of anorexia nervosa? Results from a systematic review and meta-analysis. J Clin Psychiatry. 2012;73:e757–766. 41. Schulz C, Haight R. Safety of olanzapine use in adolescents. Expert Opinion Drug Saf. 2013;12:777–782 42. Reist C, Mintz J, Alberts L. Second-generation antipsychotic exposure and metabolic-related disorders in patients with schizophrenia: An observational pharmacoepidemiology study from 1988 to 2002. J Clin Psychopharmacol. 2007;27:6–51 43. Treuer T, Hoffmann VP, Chen AK et al. Factors associated with weight gain during olanzapine treatment in patients with schizophrenia or bipolar disorder: Results from a sixmonth prospective, multinational, observational study. World J Biol Psychiatry. 2009;10(4 Pt 3):729–740. 44. Gorla K, Mathews M. Pharmacological treatment of eating disorders. Psychiatry (Edgmont). 2005;2:43–48. 45. Golden NH, Keane-Miller C, Sainani KL, et al. Higher caloric intake in hospitalized adolescents with anorexia nervosa is associated with reduced length of stay and no increased rate of refeeding syndrome. J Adolesc Health. 2013;53:573–578. 46. Steinhausen HC. Clinical guidelines for anorexia nervosa and bulimia nervosa. Eur Child Adolescent Psychiatry. 1997,6:121–128 47. Morgan HG, Purgold J, Welbourne J. Management and outcome in anorexia nervosa. A standardized prognostic study. Br J Psychiatry. 1983;143:282–287. 48. Wysowski DK, Bacsanyi J. Cisapride and fatal arrhythmia. N Engl J Med. 1996; 335:290–291. 49. Eckert ED, Goldberg SC, Halmi KA, et al. Depression in anorexia nervosa. Psychol Med. 1982;12:115–122. 50. Kripke DF. Bright light treatment reduces symptoms in older adults with non-seasonal major depression. Evid Based Ment Health. 2011;14:75. 51. Fassino S, Leombruni P, Daga G, et al. Efficacy of citalopram in anorexia nervosa: A pilot study. European Neuropsychopharmacology. 2002;12:453–459 52. Golden NH, Katzman DK, Kreipe RE, et al. Eating disorders in adolescents: Position paper of the Society for Adolescent Medicine. J Adolesc Health. 2003;33:496–503.

January 2015

Copyright © Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

59