*1.2. Evidence-Based Treatment Guidelines for Eating Disorders*

Evidence-based guidelines have been developed in several countries around the world to guide the treatment of different eating disorders, such as AN. These guidelines have the following aims [10]:


Further, guidelines can reveal gaps in the health care system [11] and inspire new paths of research. Treatment guidelines provide recommendations based on current scientific evidence. In cases where a lack of scientific evidence is available, recommendations are often provided based on expert opinion, influenced by years of clinical experience.

## **2. The German S3Guideline for the Diagnosis and Treatment of Eating Disorders**

#### *2.1. Historical Development of the S3-Guideline*

In 2000, the German Society for Psychiatry, Psychotherapy and Psychosomatics (DGPPN) published a guideline for the diagnosis and treatment of eating disorders in Germany for the first time [12]. In the same year, a guideline of the German Society of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy (DGKJP) was also published [13]. Both guidelines were developed by expert groups using informal consensus (a representative group of experts from the relevant medical society prepares a recommendation which is adopted by the board of the society, development stage one) with the aim of developing recommendations for the diagnosis and treatment of eating disorders. In the autumn of 2003, a conference of members of the German Society for Psychosomatic Medicine and Medical Psychotherapy (DGPM) and the German College of Psychosomatic Medicine (DKPM) decided to develop an evidence-based guideline for eating disorders in Germany according to development stage three (S3, based on all elements of systematic development—logic, decision and outcome analysis, evaluation of the clinical relevance of scientific studies and periodic review).

One year later, in the spring of 2004, a group composed of psychiatrists, child and adolescent psychiatrists, medical specialists in psychosomatic medicine and psychologists with expertise in eating disorders, was formed. The group included representatives of the five professional societies (DGKJP, the German Psychological Society (DGPs), DGPM, DGPPN and DKPM) that are responsible for the care of patients with eating disorders within the German health care system. In 2010, the evidence-based guideline for the diagnosis and treatment of eating disorders was published online by the Association of the Scientific Medical Societies in Germany (AWMF) [14]. The AWMF advises on matters and tasks of fundamental and interdisciplinary interest in medicine and provides, among other things, a wide range of clinical practice guidelines on its website. The AWMF is the national member for Germany in the Council for International Organisations of Medical Sciences (CIOMS) at the World

Health Organisation, Geneva. In 2011, the guideline was made available in book format [15]. Based on the scientific guideline, a patient guideline was published in 2015 [16]; this guideline, supported by the German Society for Eating Disorders (DGESS), was designed to communicate the content of the scientific guideline to patients and relatives. The patient guideline, available both online [16] and in book format [17], addresses care structures and supports communication with professional health care providers, such as the family doctor, medical or psychological psychotherapists for adults or child and adolescent psychiatrists and psychotherapists.

Over the last two years, the scientific guideline has been revised, and a second edition will be available in German at the beginning of 2019, both online [14] and in book format. An English version of the guideline is currently in preparation and will be released at a later date. The scientific guideline addresses all age groups and is available in both a short and an extended version. The thematic structure of the recent guideline largely corresponds to the first edition and includes chapters covering epidemiology, diagnostics, the therapeutic relationship, AN, Bulimia nervosa (BN), Binge eating disorder (BED), physical sequelae and methodology. The chapter 'Diagnostics' is subdivided into sections on the diagnostics of psychological and somatic symptoms. In line with DSM-5 [6], two new categories of eating disorders have been added to the revised guideline: the 'Other Specified Feeding or Eating Disorders' (OSFED), which also include the 'Night Eating Syndrome', and the 'Avoidant Restrictive Food Intake Disorder' (ARFID), which replaces the old category of 'Eating Disorders Not Otherwise Specified' (EDNOS). With regard to the therapeutic studies on AN [18], BN [19] and BED [20], meta-analyses were performed based on a systematic literature search and assignment of pre-determined quality indicators (evidence level I).

#### *2.2. Recommendation for AN—Differences between the First Version and the Revision*

Changes in treatment recommendations were based on a systematic literature search (2008–2017), in which 26 new randomised controlled trials (RCTs) on psychotherapeutic treatments, 13 new RCTs on pharmacotherapy and 2 new RCTs on nutritional management were identified [14,18]. The evidence base has considerably improved since the first version, although studies still show a large heterogeneity in terms of samples (adolescents, adults, severe and enduring AN), setting (outpatient, day hospital, inpatient), treatment phase (acute, maintenance) and outcome measures used. It should be emphasised that an improvement in study quality can be seen. In recent years, for example, studies have been published with sample sizes that allow sufficient statistical power [21–26].

Treatment recommendations were based on a network-meta-analysis (see Section 2.3), newly published RCTs, systematic reviews, or lower levels of evidence (if RCTs or systematic reviews were not available). The guideline group discussed each recommendation in light of the available evidence, clinical relevance and suitability. The most relevant changes in the revised version concerning evidence levels and recommendations are summarised in Table 1. Evidence levels were assigned using the Oxford Centre of Evidence Based Medicine criteria [27]: An evidence level of I is given if there is evidence for a specific treatment based on a systematic review (or meta-analysis) of randomised controlled trials (Ia), or one randomised controlled trial with narrow confidence interval (Ib). An evidence level of II is based on cohort-studies (IIa: systematic review, or IIb: individual cohort study). Evidence level III refers to case-control studies and evidence level IV to case-control series.

Treatment recommendations in the German treatment guideline were graded according to levels 'A', 'B', '0' and 'KKP' [28]. 'A' is the strongest recommendation, which is usually based on evidence level I (something 'is to be done'). 'B' recommendations are less strong (something 'should be done'; evidence level II) and '0' recommendations are even less explicit (something 'may be done'). 'KKP' ('clinical consensus point') stands for recommendations, which are not based on empirical research and were derived from the experience of experts (good clinical practice). Grading of recommendations was based largely on the evidence level, but also took the following criteria into account: clinical relevance of effect sizes and end points, the balance of benefits and risks, ethical considerations, patient preferences and applicability. Grading of recommendations was discussed in several consensus meetings.


#### **Table 1.** German guideline—changes in treatment recommendations for AN.

FBT, Family-Based Treatment; FPT, Focal Psychodynamic Therapy; CBT-E, Enhanced Cognitive Behaviour Therapy; MANTRA, Maudsley Model of Anorexia Nervosa Treatment for Adults; SSCM, Specialist Supportive Clinical Management.

Several key treatment recommendations did not change. They will be referred to in the comparison of evidence-based guidelines from other countries (see Section 3.2).

Up and down-grading of recommendations: Only one recommendation was downgraded. It is the recommendation for the use of low-dose neuroleptics in some cases of AN. The decision was based on the consideration that this recommendation should be followed only with caution and not as an overall clinical standard. In contrast, the recommendation not to use neuroleptics for the treatment of AN was upgraded due to an increase in evidence (systematic reviews). The same is true for the recommendation to continuously address motivation to change throughout treatment. Several studies show that motivation to change is a relevant predictor of treatment outcome. Recent high-quality trials made it possible to make specific recommendations regarding the use of specialised psychotherapeutic treatments. However, due to ethical reasons, no study compared an active treatment with untreated control groups. Therefore, it was decided that the recommendation should be classified as 'B' and not 'A'. A further recommendation was upgraded based on clinical relevance; Inpatient treatment should take place in facilities which are able to offer a specialised multimodal treatment programme. In Germany, some adult and child and adolescent psychiatric and psychosomatic hospitals are not specialised and have no experience with the treatment of patients with AN. Treatment in such facilities is, therefore, not recommended, due to high associated risks, not to mention high costs. The new guideline also includes the explicit recommendation to consider co-morbidity in patients with AN. Co-morbid conditions like borderline-personality disorder or post-traumatic stress disorder, for example, might require changes in treatment planning and prioritisation of therapy goals. Although empirical evidence is scarce, a recommendation for a stabilisation phase as a final phase in inpatient treatment was included, as relapse after discharge is common [29–31], and the transition from one service level to another service level (especially to a level with less supervision and support) is a major challenge for patients with AN. Finally, there was new empirical evidence suggesting that a short inpatient stay for weight stabilisation followed by day hospital treatment is as effective as long-term inpatient treatment for children and adolescents with AN, providing there is continuity in the therapists that are responsible and if there is sufficient support by family members [23].

#### *2.3. Network-Meta-Analysis*

Based on the systematic literature search (see Section 2.2), a network-meta-analysis was conducted to answer the following question: What is the comparable effectiveness of different psychotherapeutic treatments for AN? Additionally, two further questions were addressed using standardised mean change statistics: What is the amount of weight gain that can be expected in different treatment settings? And: What is the amount of weight gain that can be expected in adolescents vs. adults?

Predefined inclusion and exclusion criteria were used to select the studies. Each study was rated by two independent researchers and additionally assessed for quality [18]. For more details on data analysis see [18].

Network-meta-analysis: 18 randomised controlled studies met inclusion criteria for the data-analysis. Ten studies were on adolescents (625 patients), and 8 studies were on adults (622 patients). No treatment approach was found to be superior. However, there were several limitations to the analysis and interpretation of results. All studies compared active treatments with each other, with no study including an untreated control group. Only a few comparisons were replicated. Furthermore, the majority of studies on adolescents evaluated family-based treatment approaches mostly by the same group of researchers, while interventions in adults were almost exclusively on an individual basis. The manualised treatment approaches that were evaluated in high quality trials comprise the Maudsley Model of Anorexia Nervosa Treatment for Adults (MANTRA) [25], Focal Psychodynamic Therapy (FPT) [26,32,33], Enhanced Cognitive Behaviour Therapy (CBT-E) [26,34,35], Specialist Supportive Clinical Management (SSCM) for adults [25,35,36], and family-based treatment (FBT) for adolescents [21,22,24].

Standardised mean change statistics (SCM): Analyses were conducted with 38 studies (1164 patients). Seventeen of these studies were naturalistic studies, and four studies were on adolescents (350 patients). For a course of up to 27 weeks, significantly higher weight gains can be expected in inpatient treatment compared to outpatient treatment (for adults: mean weight gain of 537 g/week in inpatient treatment vs. 105 g/week in outpatient treatment; for adolescents: mean weight gain of 615 g/week in inpatient treatment vs. 192 g/week in outpatient treatment). The estimated effect sizes for weight gain in adolescents were significantly higher compared to adults (in RCTs: SMC = 1.97 vs. 1.02, in naturalistic studies SMC = 1.84 vs. 1.42, respectively).

In sum, there are several existing manualised psychotherapeutic treatments for AN, which can be considered evidence-based and effective. However, there is a need for replication studies. There are differences regarding treatment response and most suitable treatment approach in adult versus adolescent patients.
