In anorexia nervosa (AN) it still remains to be clarified, which psychiatric symptoms are the direct consequence of malnutrition and adaptation to starvation and which are not. There is clinical consensus that depression/anxiety and cognitive impairments in AN mainly are sequelae to the malnutrition. However, this consensus is largely based on experimental starvation studies of healthy subjects back in 1940s and from famine- and food programs in the third world, whereas evidence from studies on AN is lacking.

The main objective of the study is in the detail to elucidate the short-term changes in the psychopathological profile, depression, anxiety, and cognitive functions in relations to intensive nutritional rehabilitation with weight gain of 10-30% in a specialized medical stabilization unit. Secondarily, it is examined whether cortisol levels are associated with depression/anxiety symptomatology and cognitive impairments. The hypothesis is that an improvement in nutritional status over a short time leads to clinically significant improvements in psychopathology and cognitive functions, which remain 2-3 months after discharge, making the patients more accessible to psychotherapy.

Anorexia nervosa (AN) is a syndrome characterized by a distorted body image and morbid fear of obesity. Restrictive eating and compensatory behavior in the form of excessive exercising or vomiting result in malnutrition. Repeated revisions of the diagnostic criteria in recent decades have by definition resulted in an increase in prevalence. The disease has the highest mortality among all psychiatric disorders. The majority of deaths is related to malnutrition and its consequences. There is no effective evidence-based treatment. At least 25% remain chronically undernourished and have a long course with prolonged hospitalizations in both psychiatric and medical units, social psychiatric institutions and - support, early retirement and premature death. The etiology remains unknown and it is unclear which of the psychological symptoms that are a consequence of malnutrition, and which ones are premorbid or comorbid.

Depression and anxiety are co-existing and familiar features of patients with AN and can occur in such a degree that it is termed comorbidity. Antidepressant therapy has not shown to have any therapeutic effect of AN. Varying degrees of cognitive impairment are also well known feature of the AN, but in the literature it is only scanty investigated and psychometric analyzed.

In the early stage of the disease, starvation and weight loss can be interpreted as a coping strategy that relieves anxiety and depressive symptoms. If this coping strategy works, then the effect may decreases in line with the severity of malnutrition. There is almost a clinical consensus that depression / anxiety and cognitive impairment in AN is sequelae to malnutrition. But, it is a consensus that rely mainly on experimental studies of healthy subjects back in the postwar years observations from famine and from food programs in the 3rd World. Knowledge of the relationship between psychopathology and malnutrition in AN is still limited to a few and small studies with contradictory conclusions. In a recent study of hospitalized patients, no correlations between body mass index (BMI) and selected psychometric variables could be detected, with the essential limitation that it was a cross-sectional study of a population of patients with a narrow BMI range. To our knowledge, only one longitudinal study on changes in BMI and depressive symptoms have been published. In that study, mean weight gain was a significant predictor of reduced score in several of the "specific" eating disorder symptoms, such as concerns about diet, while weight gain was not a predictor of concern about body nor for depressive symptoms. To our knowledge, no longitudinal re-nutrition studies of the relationship between weight gain and anxiety or cognitive functions has been published.

A consequence of starvation stress is that all the endocrine axes are altered in AN, leading to protein- and energy preserving adaptation. Thus, in particular, increased levels of cortisol is well described in both AN and in patients with primary affective disorders. Depression/anxiety and cognitive impairment is also a well-described side effect of pharmacological treatment with corticosteroids and is well-known symptoms in excess endogenous cortisol production (mb. Cushing). Moreover, neuroanatomical animal studies indicate that cortisol per se has neurotoxic properties and may play a pathogenic role in affective disorders and anxiety. There is preliminary experimental evidence that pharmacological cortisol synthesis inhibition may have therapeutic efficacy in depression. However, so far this is not investigated in AN. In one study, it was found that cortisol levels correlated positively with anxiety and depression.

Center for Eating Disorders at Odense University Hospital is one of three highly specialized national centers in Denmark. Within the center, there is a formalized collaboration between the psychiatric and somatic units. Patients with life-threatening weight loss are primarily hospitalized in the Nutrition section for somatic stabilization and weight gain typically, 10-40%, and then either transferred to the psychiatric department or discharged for two-track psychiatric / somatic outpatient treatment. The Nutrition section receives severely ill patients from the nation. The median BMI of 84 admissions in year 2013 was 13.8 (the range 7.8 - 25.8). Internationally, there are only few somatic units of similar specialization and patient volume, which constitute the basis for studying the effect of intensive re-nutrition per se. The center's organization and patient population is described in several observational and intervention studies.

International and national guidelines recommend that treatment of AN should be interdisciplinary and double-track psychiatric - somatic. But, it is not clear how far and how fast nutritional rehabilitation may have beneficial effects on depression, anxiety and cognitive impairment in patients with severe AN.

Inclusion criteria:

- Fulfilling Diagnostic and Statistical Manual (DSM)-5 criteria for AN.

Exclusion Criteria:

• Patients who can not fill Danish-language questionnaires.
• Patients who are primarily hospitalized a few days for water electrolyte correction only.
• Patients with active drug abuse.
• Patients with co-morbid schizophrenia.
• Patients where the psychopharmacological treatment qualitative is changed during the study period.

Patients with severe AN referred to a specialized unit.

Data will be entered in Odense Patient data Explorative Network (OPEN).