Introduction
Eating disorders (EDs) represent a group of persistent and potentially fatal psychiatric conditions that rank among the top-10 leading causes of disability in young adults.Reference Simon, Schmidt and Pilling1, Reference Hudson and Pope2 Typically, clinical onset is during infancy/adolescence, and less frequently in early adulthood. The estimated lifetime prevalence of EDs is now considered as ranging between 0.3% and 0.9% for anorexia nervosa (AN) and up to 10% for other conditions.Reference Hoek and Van Hoeken3–Reference Machado, Machado, Gonçalves and Hoek7 EDs are characterized by heterogeneous clinical presentations and include anorexia nervosa restricting type (AN-R), anorexia binge-eating/purging type (AN-BP), bulimia nervosa (BN), and binge-eating disorder (BED), and also encompass other atypical or subthreshold forms. Moreover, even if AN and BN are typically regarded as discrete diagnostic entities, they share several psychopathological elements, and patients frequently cross the diagnostic categories.Reference Castellini, Lo Sauro and Mannucci8–Reference Strober, Freeman and Morrell11 Not surprisingly, it has been reported that in 20–50% of cases, AN and BN might occur together across the lifespan.Reference Eddy, Dorer, Franko, Tahilani, Thompson-Brenner and Herzog12
Anorexia nervosa (AN) is the most severe ED. It affects about 1:200 women and 1:2000 men in western countries.Reference Hoek and van Hoeken13 The etiology of AN is complex, and several factors, such as genetics and personality traits, may contribute to its development and maintenance. In addition, environment, such as sociocultural pressure toward thinness, is likely to act as a trigger during the vulnerable period of pubertal development. The increase in satisfaction gained from bodily and dietary control is often reinforced in societies where a number of individuals are facing problems of overweight and obesity. As a consequence, AN patients have a tendency to conceal their problems, and to seek for professional help only when the long-lasting malnutrition produces severe physical complications, including osteoporosis, gastrointestinal and cardiac failures, liver damage, or electrolyte disturbances.
According to the National Institute for Clinical Excellence Guidelines (NICE),14 and also according to the Italian Guidelines,Reference Brambilla and Tortorella15 most patients with AN should be managed on an outpatient basis using psychological treatment methods. Hospitalization has to be considered when there is substantial medical or suicidal risk, or after failure to improve despite an adequate course of psychotherapy. Inpatient programs provide structured regimens focused on refeeding and weight gain in combination with broader psychosocial interventions. Whenever possible, hospitalization should occur within or near the patient’s own community and followed by a minimum of 12 months of outpatient treatment. First-choice psychological treatments for AN include cognitive behavior therapy (CBT), interpersonal psychotherapy (IPT), psychodynamic therapy, family interventions, psychosocial interventions based on addiction models, support groups, Internet-based support, and nonverbal therapeutic methods. Psychopharmacological treatments represent only a second choice for those patients who are resistant to psychotherapy.14–16
In any case, a summary of 68 studies dealing with psychotherapeutic and pharmacological strategies in AN published before 1989 with a follow-up length ranging between 1 and 33 years reported that only 43% of patients recovered completely, 36% improved, 20% developed a chronic form, and 5% died from physical complications and/or suicide.Reference Steinhausen17 Subsequent long-term observations reported on a standardized mortality ratio (SMR), (defined as the ratio of observed deaths in the study population to expected deaths in population of origin) of 5.86 (95% CI: 4.17-8.26) with a mean follow-up period of 14 years.Reference Smink, van Hoeken and Hoeck18 Moreover, clinical experiences would indicate that a large number of patients show unsatisfactory response to all available treatments.
Clinical trials carried out regarding AN up until now have focused on a wide range of pharmacological compounds, but unfortunately, data on their effectiveness are still limited. As a consequence, currently there is no proven or unequivocal treatment for this disorder. However, it may be wrong to consider the modest results obtained as “absence of evidence.” The present article aims at providing an exhaustive review of current knowledge on psychopharmacological options for adult patients with AN. Studies only on children or adolescents were not included, considering that the vast majority of drugs are not permitted for patients younger than 18 years. Conversely, studies including both patients younger and older than 18 years were included.
Methods
The following criteria were used to select studies assessing psychopharmacological options for AN patients: (a) inclusion in at least 1 of 2 databases, MEDLINE and PsycINFO, from January 1966 to January 2014; (b) original articles published as full papers, brief reports, case reports, or case series; (c) original articles not specifically focused on special populations (namely, children and adolescents); (d) original articles including patients identified as suffering from some form of AN based on international diagnostic standards, such as the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), DSM Fourth Edition Text Revision (DSM-IV-TR), or International Classification of Diseases (ICD)-9 or −10; (e) studies published in the English language. The following keywords were used: [eating disorders], [anorexia], [treatment], [resistance], [refractory]. A specific form was designed for data extraction, including (a) authors; (b) therapeutic setting (inpatient or outpatient) and AN diagnosis; (c) number of subjects; (d) type of treatment; (e) type of study; and (g) outcomes. All of the studies reviewed are reported in Table 1.
Table 1 Studies on psychopharmacological treatments available for anorexia
List of Abbreviations: HSCL: The Hopkins Symptom Checklist; HSCL-90: The Hopkins Symptom Checklist 90–Revised; GAAQ: Goldberg Anorectic Attitude Questionnaire; PRS: Psychiatric Rating Scale; ABOS: Anorectic Behaviors Scale for Inpatients Observation; EAT: Eating Attitudes Test; BAT: Body Attitudes Test; SADS-C: Schedule for Affective Disorders & Schizophrenia; CGI: Clinical Global Impression; CGI-I: Clinical Global Impression–Improvement; CGI-S: Clinical Global Impression–Symptoms; HAM-D: Hamilton Depression Scale; BDI: Beck Depression Inventory; BDS: Beck Depression Score; SEI: Self Esteem Inventory; ABS: Anorexic Behavior Scale; EDE: Eating Disorders Examination; EDI: Eating Disorder Inventory; BIT: Body Image Test; BDS: Beck Depression Score; BSQ: Body Shape Questionnaire; TCI: Temperament and Character Inventory; SCL-90: Symptom Checklist–90-R; Y-BOCS: Yale–Brown Obsessive Compulsive Scale; YBC-EDS: Yale–Brown–Cornell Eating Disorder Scale; YBOCS-ED: Yale–Brown Obsessive Compulsive Scale for Eating Disorder; STAI: State-Trait Anxiety Inventory; HDRS: Hamilton Depression Rating Scale; HARS: Hamilton Anxiety Rating Scale; STAXI: State-Trait Anger Expression Inventory; TESS: Traumatic Exposure Severity Scale; BAI: Beck Anxiety Inventory; RSE: Rosenberg Self-Esteem; Q-LES-Q: Quality of Life Enjoyment and Satisfaction Questionnaire; SAS: Sympson-Angus Scale; CDR: Cognitive Drug Research; PANSS: Positive and Negative Symptoms Score; CESD: Center for Epidemiologic Studies Depression Scale; MASQ: Mood and Anxiety Symptoms Questionnaire; PWI: Personal Wellbeing Index; MADS: Managing Affect and Differences Scale; SCID-I: Structured Clinical Interview for DSM-IV-TR (Axis I Disorders); SCID-II: Structured Clinical Interview for DSM-IV-TR (Axis II Disorders); BPRS: Brief Psychiatric Rating Scale; PAI: Personality Assessment Inventory; MADRS: Montgomery–Asberg Depression Rating Scale; SAPS: Scale for the Assessment of Positive Symptoms–Delusion Subscale; AIMS: Abnormal Involuntary Movement Scale; EDI-SC: Eating Disorder Inventory (2)–Symptom Checklist; LIFE II BEI: Eating Disorder Interview Based on Long Interval Follow up Evaluation; EDI-2: Eating Disorder Inventory-2; EDE-12: Eating Disorder Examination–12th Edition; STAI-Y: Speilberger State-Trait Anxiety Inventory; VAS: Visual Analog Scale; PBO: Placebo; RCT: Research Clinical Trial; CBT: Cognitive-Behavioral Therapy; AN-R: Anorexia Restricter; AN-BP: Anorexia Bingeing-Purging.
Results
A total of 75 publications was found. Thirty-four references were excluded following the selection criteria (Figure 1). The full text of 41 papers retrieved from PubMed and PsycINFO searches was selected and analyzed. The main characteristics of the 41 studies are listed in Table 1. In view of the paucity of reports, we did not apply rigorous quality criteria to the experimental designs of studies, such as substantial subject number, random assignment, adequate control subjects, reliable methods of assessment, or long-term follow-up. The heterogeneity of sample compositions, treatment, settings of treatment, and analyzed predictors did not permit us to carry out a meta-analysis of all of the available studies.
Figure 1 Flow chart summarizing the procedure for selecting studies for review.
Seventeen (41.5%) studies used a randomized controlled trial (RCT) methodology, while the others were open (n=9, 22.0%), case series and case reports (n=12, 29.3%), or retrospective observations (n=2, 4.9%). One trial was conducted with a single-blind randomized design (2.4%).
The numbers of subjects included in the studies varied widely, between 1 and 93 (mean±SD: 22.3±22.4). Three studies also included men (7.3%) for a total of 4 patients, therefore making a comparison based on gender impossible. The size of the samples and the diagnosis/gender composition depended on the characteristics of the setting (inpatient/outpatient) and on the duration of observation, whereas the power of the study was not considered in the majority of published papers. Results were statistically nonsignificant not because the hypothesis being tested was untrue or clinically nonsignificant but because the sample size was too small.
After differentiating by setting, 25 studies were conducted on inpatients (61.0%), 15 on outpatients (36.6%), while one case did not specify the therapeutic setting of the patients. The total number of inpatients included was 594 (mean+SD: 23.7±23.2, range: 1–93) that of outpatients 321 (mean+SD: 21.4±21.7; range: 1–81).
The age (mean ± SD, years) of patients was mentioned in 35 out of a total of 41 studies (85.3%), and was 27.2±8.1 years (range: 18.4–50.0). The age of inpatients was significantly lower than that of outpatients (24.8±4.2 vs 29.1±9.5, t-test: p<.018).
The illness duration (mean±SD, years) for the overall settings (available data for 19 out of 41 observations, 46.3%) was 7.9±7.1 (range: 2.9–35.0), that of inpatients was 6.8±3.4 (range: 2.9–14.1), and that of outpatients was 5.5±1.4 (range: 3–7).
With regard to DSM-IV or DSM Third Edition Revised (DSM-III-R) diagnoses, 18 studies did not specify the AN subtype (43.9%), 9 (22.0%) included only patients with AN-R, 2 (4.9%) included those with AN-BP, and 11 (26.8%) included both AN-R and AN-BP patients. One study (2.4%) included patients with AN and BN.
Out of a total of 41 observations (70.7%), 29 investigated Axis I comorbidity, with 6 studies (14.6%) excluding patients with Axis I mood or psychotic disorders. OCD spectrum comorbidity was diagnosed in 9 studies (22.0%). No information was provided in 12 studies (29.3%). Only four studies (9.8%) investigated Axis II comorbidities.
Twenty-four studies (58.6%) used ED-related inventories, while 17 (41.4%) utilized no rating scales. Details on rating scales administered are summarized in Table 1. Body mass index (BMI) at intake was available in 21 studies (51.2%). Mean BMI at baseline was 14.9±1.9 kg/m2 (range: 9.8–17.9), and at the end 17.1±1.5 kg/m2 (range: 14.7–20).
Discussion
The results of the present review confirm the paucity of empirical evidence on the available pharmacological treatments of adult AN. Therefore, challenges to the identification of evidence-based treatments for AN are discouraging. Available studies are affected by several methodological limitations, so that the overall strength of the available evidence is poor, with a lack of consensus and inconclusive outcomes. Most studies were conducted at a single center. None of the pharmacological options tested, including antidepressants, antipsychotics, or mood stabilizers, achieved relevant treatment goals. Sample sizes were small, and dropout rates were often high. The few results that favored one form of treatment over another were not replicated. As a consequence, impressions about “positive” and “negative” findings for AN were shaped by single studies with a few participants, and were often extrapolated across patient groups of different age, illness duration, and severity. Here the available evidence is summarized by drug class.
Antidepressants
The biological and psychopathological similarities between AN and depression led to the early clinical trials with tricyclic antidepressants (TCAs), such as amitryptylineReference Biederman, Herzog and Rivinus19, Reference Halmi, Eckert, LaDu and Cohen20 and clomipramine.Reference Lacey and Crisp21, Reference Crisp, Lacey and Crutchfield22 However, TCAs are no longer studied or recommended for AN, mainly because they may induce fatal arrhythmias in patients with low body weight (especially in younger patients).
Selective serotonin reuptake inhibitors (SSRIs) replaced TCAs in the 1980s. Again, the rationale for the use of SSRIs was related to the similarities between AN, depression, and OCD. According to the spectrum model proposed by Hollander et al,Reference Hollander, Kim, Braun, Simeon and Zohar23 AN restricting subtype could be included in the obsessive-compulsive spectrumReference Phillips, Friedman, Stein and Craske24 and underpinned, perhaps, by a serotoninergic dysfunction.Reference Ramacciotti, Coli, Paoli, Marazziti and Dell’Osso25 However, to date, there is no clear evidence supporting the use of SSRIs in AN. Studies on fluoxetine provided controversial results. The first trial was a small, open study in 6 patients with chronic, refractory AN.Reference Gwirtsman, Guze, Yager and Gainsley26 The authors observed an improvement of depressive symptoms and a significant weight gain. Patients seemed to tolerate fluoxetine even at the higher dosages. Conversely, no differences were found between fluoxetine and placebo in 2 studies conducted in AN-R inpatients.Reference Attia, Haiman, Walsh and Flater27, Reference Barbarich, McConaha and Halmi28 In a third study, patients with AN-R were randomly assigned to fluoxetine or placebo after a period of hospitalization and weight restoration, and followed up for 1 year.Reference Kaye, Nagata and Weltzin29 Patients receiving fluoxetine showed a significantly lower rate of relapse than those randomized to placebo. However, the limitations of this study included the small sample size (13 completers) and the lack of standardized psychological treatment during the trial. Negative results were found in a larger clinical trial, with 93 weight-restored outpatients already in treatment with CBT randomized to fluoxetine or to placebo.Reference Walsh, Kaplan and Attia30 No differences in relapse rate emerged during the 1-year follow-up.
Citalopram was compared to a waiting-list condition (WL) in patients affected by AN-R (participants in WL received no active treatment during the trial; it was anticipated that they could receive one after the end of the study).Reference Fassino, Leombruni and Daga31 Citalopram improved depression, obsessive-compulsive symptoms, impulsiveness, and trait anger. Weight gain was similar in both groups.
Mirtazapine showed its effectiveness in an adult patient, in which weight-gain maintenance and mood improvement were observed in a 9-month follow-up.Reference Safer, Darcy and Lock32 The same research group reported the efficacy of high doses (180 mg/day) of duloxetine in the management of treatment-refractory symptoms in a patient affected by severe depression, OCD, and comorbid AN-BP, with a significant improvement of affective, obsessive, and eating-refractory symptomatology.Reference Safer and Arnow33
In conclusion, available data on antidepressants demonstrated their limited effectiveness in AN. However, they might improve the anxiety, irritability, mood lability, and depressive symptoms that often complicate the course of AN.
Antipsychotics
In the past, the rationale for the use of antipsychotics in AN was mainly linked to the psychomotor hyperactivity control (with the induction of weight gain), or to the reward system regulation.Reference Vandereycken34 Several core symptoms of AN, including body image disturbance and fear of weight gain, have been closely related to an underlying dopaminergic dysregulation.Reference Ramacciotti, Dell’Osso and Paoli35 A recent study postulated the existence of an altered expression of dopaminergic genes among patients suffering from EDs.Reference Frieling, Römer and Scholz36 This provided a more robust rationale for the use of antipsychotics, especially when fears of weight gain and body image disturbance might reach a delusional level.
Data on first-generation antipsychotics were controversial.Reference Cassano, Miniati and Pini37–Reference Vandereycken and Pierloot41 Results regarding the safety and efficacy of haloperidol on delusional body image alteration were promising, but were limited by the small samples size. To our knowledge, only 2 observations were available concerning haloperidol as adjunctive treatment for AN-R. The first was a mid-term naturalistic study on a sample of 13 outpatients (mean BMI of 15.6 kg/m2)Reference Cassano, Miniati and Pini37; the second one was a chart-review of 9 inpatients (BMI<13 kg/m2).Reference Mauri, Miniati, Mariani, Ciberti and Dell’Osso38 Chlorpromazine is no longer utilized for AN because of its severe adverse effects, including seizures. These preliminary studies suggested that low doses of haloperidol might be effective as an adjunctive treatment for patients with severe, treatment resistant AN-R and delusional body image alteration.
There was inconsistent evidence from 2 trials on the efficacy of sulpirideReference Vandereycken34 and pimozide.Reference Vandereycken and Pierloot41 In a third single-blind study, amisulpiride (50 mg/day) was compared with fluoxetine and clomipramine in a small sample (n=35) of hospitalized patients with AN who attended a 12-week weight restoration program. Those taking amisulpiride showed a significantly greater weight gain, while the other clinical features did not significantly differ between groups.Reference Ruggiero, Laini and Mauri40
More evidence on second-generation antipsychotics is now available. Olanzapine is the most studied antipsychotic, with 8 trials in adult AN patients (4 compared to placebo, 1 to chlorpromazine, 2 open label, and 1 retrospective study)Reference Powers, Santana and Bannon42–Reference Brambilla, Garcia and Fassino49 and 4 case reports involving adult patients.Reference Hansen50–Reference Wang, Chou and Shiah53 The first 10-week, open-label trial examined the efficacy and safety of olanzapine 10 mg daily in 18 patients (2 male) with AN, with 10 patients gaining weight.Reference Powers, Santana and Bannon42 Again, 17 AN women received olanzapine 2.5–7.5 mg daily (mean 4.7) in an open-label fashion, and were evaluated at baseline and every 2 weeks. Twelve patients completed the 6-week study period, and showed significant improvements in weight, and in depression and anxiety signs and symptoms.Reference Barbarich, McConaha and Gaskill43 Another open-label study comparing olanzapine and chlorpromazine in the treatment of 15 AN patients showed no significant differences in the average weight gain between the 2 groups, although there was a statistically significant reduction in ruminative thinking in the olanzapine group vs the chlorpromazine group (p<.01).Reference Mondraty, Birmingham, Touyz, Sundakov, Chapman and Beumont44 The effect of olanzapine on behaviors associated with AN was evaluated in a retrospective study on 18 inpatients (mean age 22 years).Reference Malina, Gaskill and McConaha45 The average duration of treatment was 17 weeks (range 3–70). Improvements were observed in the frequency of obsessive thoughts about body image and anxiety before and during meals (p<0.001), and in the ability or desire to eat (p<0.001). Moreover, patients reported being “less upset about weight gain” (p=0.002), “less upset in stressful situations” (p<0.01), and “better able to fall asleep at night” (p<0.01). In any case, all these trials suffer from several limitations, such as the open-label design, the brief duration, the small sample sizes, the low completion rates, and the presence of confounding factors, such as concomitant treatments (benzodiazepines [BDZs], SSRIs, and psychotherapies).
In the first study vs placebo, a survival analysis was conducted comparing treatment conditions in time to achievement of target BMI (18.5 kg/m2).Reference Bissada, Tasca, Barber and Bradwejn46 Of the total sample, 55.6% of patients receiving placebo and 87.5% of patients receiving olanzapine achieved weight restoration (Mantel-Cox test: χ2=5.42, df=1, p=0.02). Psychological outcomes according to treatment condition were also investigated and indicated that olanzapine resulted in greater positive change in obsession scores than placebo (β11=4.20; t=2.37, df=55, p=0.02). No differences in compulsion scores were demonstrated (β11=0.60; t=0.45, df=55, p=0.70).Reference Bissada, Tasca, Barber and Bradwejn46 The second double-blind study was conducted on AN patients≥16 years. Participants were free of other psychiatric medications for at least 4 weeks prior to study enrollment, with the exception of SSRIs or SNRIs that were permitted if doses had not changed for 4 weeks prior to study enrollment. Participants did not receive individual psychotherapies. A total of 23 patients was randomized either to olanzapine or placebo for 8 weeks. End-of-treatment BMI was significantly higher in the olanzapine group (p<.018). Psychological symptoms improved in both groups, but there were no statistically significant differences.Reference Attia, Kaplan and Walsh47
In addition, 2 double-blind randomized studies have been published on psychobiological effects of olanzapine in AN patients.Reference Brambilla, Monteleone and Maj48, Reference Brambilla, Garcia and Fassino49 In the first one, one-half of the AN patients (including both the AN-R and the AN-BP subtypes) received a combined treatment of CBT and double-blinded olanzapine, whereas the other half received CBT+placebo. Homovanillic acid (HVA) blood concentrations for dopamine secretion were monitored at baseline and then monthly during the trial. BMI increased significantly in both treatment groups, with no difference between the two. There were also no significant differences between the groups regarding the Eating Disorder Inventory-2 (EDI-2) scores. HVA plasma concentrations did not change in the CBT+placebo patients, whereas they increased significantly in the CBT+olanzapine group.Reference Brambilla, Monteleone and Maj48 In the second study, 20 AN patients received CBT for 3 months and a programmed nutritional rehabilitation, combined with olanzapine vs placebo (2.5 mg for 1 month and 5 mg for 2 months). BMI, leptin, and ghrelin plasma values were monitored at baseline and then monthly for 3 months. BMI increased significantly, but not differently in both treatment groups. Leptin and ghrelin secretions did not change during the course of the treatments, with no correlations with BMI values.Reference Brambilla, Garcia and Fassino49
Controversial results came from 2 small open studies with quetiapine. In the first one, quetiapine was effective on both BMI restoration and on the “restraint subscale” of the Eating Disorder Examination (EDE).Reference Bosanac, Kurlender and Norman54 Conversely, in the second study, no significant changes in BMI were found, and the drug was effective in reducing anxiety symptoms and depressive symptoms comorbid with AN.Reference Powers, Bannon, Eubanks and McCormick55
In a long-term study (from 4 months to 3 years), the add-on of aripiprazole to antidepressants provoked a significant reduction in eating-specific anxiety and obsessive thoughts about food, weight, and body image.Reference Trunko, Schwartz, Duvvuri and Kaye56
Taken as a whole, findings on antipsychotics demonstrate that, except perhaps haloperidol and olanzapine, they produce no significant effect on the core symptoms of AN, such as dysmorphophobia. Moreover, BMI does not seem to be significantly influenced by atypical antipsychotics, while suggesting that the simple increase of appetite is not enough to improve the clinical picture.
Lithium
The rationale for using lithium in patients with AN resistant to treatment was questionable since its first proposal, and mainly related to the observations of its induction of weight gain. In 1977, BarcaiReference Barcai57 carried out a pilot report on lithium in adult AN. Two patients suffering from AN for many years “who wished to gain weight” were treated successfully with lithium. One patient gained 12 kg and the other 9 kg within 6 weeks. The weight gain was maintained for 1 year of follow-up with lithium. Subsequently, a placebo-controlled trial with lithium was carried out in 16 patients.Reference Gross, Ebert, Faden, Goldberg, Nee and Kaye58 There were significant differences in weight gain at weeks 3 and 4 (difference: 3.9 kg), but the sample size was small and the follow-up duration was short. In a review of clinical trials with lithium, positive results were described in a group of disorders with underlying affective dysregulation, including premenstrual tension and AN.Reference Yung59 However, the use of lithium is no longer recommended in AN, even for patients with severe and resistant forms. Sodium and fluid depletion may lead to reduced lithium clearance, resulting in an increased potential toxicity.
Other pharmacological treatments
Reports on other pharmacological treatments of AN were anecdotal. Oral clonidine was administered to a small sample (n=4) of treatment-resistant patients, with negative results.Reference Casper, Schlemmer and Javaid60 Goldberg et al Reference Goldberg, Halmi, Eckert, Casper and Davis61 used cyproheptadine (CYP), in a randomized placebo-controlled trial with 4 arms (n=81). No clinically significant effect on weight gain with CYP was found, even at 32 mg/day. As already described for studies on TCAs, cyproheptadine (32 mg/day) was compared to amitriptyline (160 mg/day) and placebo in a sample of 72 AN patients aged 13–36 years.Reference Halmi, Eckert, LaDu and Cohen20 Cyproheptadine decreased the length of time to achieve the goal of weight gain.
Conclusion
Several studies have addressed the problematic issue of the psychopharmacological management of adult AN, but only a few have offered valid options alternative to psychotherapy.
The findings of our review identified shortcomings in study design, including treatment-specific biases arising from small sample sizes, differences in study protocol, research conducted in single centers, and clinical rather than statistical interpretation of results. These limitations derive from the difficulties in conducting research studies with the AN population, including diagnostic heterogeneity, ethical issues in clinically relevant treatment, and challenges in the organization of a mid-term/long-term follow-up. Taken as a whole, treatment responses are unsatisfactory, raising the question of whether treatment resistance is a condition worthy of further clinical or biological exploration. Moreover, the merits of randomized controlled trials in this specific field are debatable, considering that, despite an extraordinary expenditure of efforts, the majority of trials produced equivocal findings.
Of the small amount of evidence that has emerged from the present review is that antipsychotics, such as olanzapine or haloperidol, might be useful in treatment-resistant AN, but only when severe or delusional dysmorphic features are present.
We believe that, above all, the disappointing findings of treatment research highlight the need for a better understanding of AN psychopathological features. Randomized controlled trials failed because several individuals with AN rejected treatment, dropped out prematurely, and sustained few behavioral changes in the absence of external concurrent factors. These outcomes seemed to be linked to patients’ attitudes about their symptoms, which often included the conviction that thinness and restraint are more important than recovery. The influence of such over-valued ideas might help to explain why this longstanding eating disorder has remained “impressively resistant to a wide range of interventions.”Reference Walsh62 According to our proposal, the search for more effective forms of pharmacotherapy should begin with a closer examination of the factors that make AN specifically difficult to study and to treat. We believe that it would be helpful to refine more accurate definitions of AN phenotypes and their treatment resistance. It is unclear if a distinction can be made between such resistance and the natural history of the disease, at least in its current clinical manifestations with available treatments. As a consequence, a fundamental recommendation concerning the profoundly important clinical challenge of improving both short- andlong-term treatments of AN patients is to employ designs of therapeutic trials with instruments that are able to detect the complete range of clinical manifestations of this polymorphic condition. Future studies should offer interventions that are better matched to the well-studied features of this disorder.
Disclosures
The authors do not have anything to disclose.
