Menatetrenone (MED) (Vitamin K2)

Administration of MED (vitamin K2) in AN among Japanese women over a nine-month period reduced bone loss, but there was no increase in BMD [181].

Evidence-based data from these recent comprehensive studies suggests that the safest and most effective strategy to protect and improve bone density in AN and prevent fractures among adolescent and premenopausal women is restoration of weight with the resumption of menses. The most promising available medications include 17 ß estradiol replacement (such as the transdermal estradiol patch) for adolescents and bisphosphonates for adults.

#### Vitamin D and Calcium Supplementation

Low levels of vitamin D and inadequate calcium intake are associated with increased fracture risk and low BMD. In addition, D3 hypovitaminosis could be responsible for the lack of inflammatory response and depressive symptoms among patients with long-term eating disorders. Calcium intake among adolescent AN patients has been described as comparable to that among controls, partly as a result of supplements [182]. While their intake and the bioavailability of oral ergocalciferol among young AN patients was similar to that of healthy controls, AN patients have lower serum levels of 25 and 1,25OH-Vitamin D [183]. In addition, patients who have lower serum levels of vitamin D (<20 ng/mL) have lower hip BMD [184]. After weight gain, the spine BMD increase was greater in the group of patients with higher serum vitamin D levels (≥30 ng/mL) [185], supporting the use of oral vitamin D supplements to obtain sufficient serum levels during weight gain. No RCT prospective trial has been performed to evaluate the efficacy of calcium and vitamin D supplements alone on BMD among AN patients. Nevertheless, given the impact of calcium and vitamin D on bones, although the efficacy of calcium and vitamin D supplementation has been poorly evaluated in AN, we recommend a total intake of calcium of approximately 1000–1200 mg and vitamin D supplements if serum levels of vitamin D are insufficient or if there is a secondary hyperparathyroidism [186]. Since oral calcium supplements can be associated with an increased risk of incident coronary atherosclerosis, a calcium-rich diet should be privileged if possible among AN patients [187].

#### **4. Discussion**

The original aim of this multidisciplinary overview was to summarize all the literature published about the use of medication in the psychiatric, somatic and nutritional aspects in AN.

Evidence based on the efficacy of medication in anorexia nervosa is scarce whether for the psychological, somatic or nutritional sphere. The evidence base is sparse, as the literature reports mainly case reports, cases series, open studies and some RTCs. In addition, for the RCTs, methodologies have numerous failings such as the heterogeneity of study designs, research methods, population samples, and intervention modalities [6,37,44,45,50]. The aims of the published research were initially to find medication that would cure AN by overcoming the key resistant symptom, the need for weight gain. Many drugs have been tested in AN on the rationale of their side effect in terms of weight gain in other disorders (for example lithium, antihistamines, cannabinoids, etc.) or their action on clinical manifestations observed in AN (depression, anxiety, obsessions, hyperactivity). In parallel, some psychotropic drugs have been tested in connection with hypotheses of neurobiological etiopathogenic neuro-transmitter involvement in the development or maintenance of AN (for example noradrenalin and serotonin for antidepressants, dopamine for antipsychotics). Recently, new neuro-hormones involved in the regulation of dimensions altered in AN, such as appetite regulation (ghrelin) or social interactions (oxytocin) have been evaluated. However, oxytocin and ghrelin have not proved their efficacy nor their safety in AN. In relation to its anorexigen effect, oxytocin needs more evaluation in AN setting [188]. These neuro-hormones should not be used in routine care practice.

#### *4.1. Somatic and Nutritional Aspects*

Concerning somatic aspects, if weight gain is a crucial step to correct the majority of adaptive and functional changes resulting from undernutrition, somatic drugs have been explored from the late 1990s. The evidence of a higher risk of osteoporosis and spontaneous fracture among AN subjects with amenorrhea [17,19] raised the question of the use of bisphosphonate or oestrogen replacement to treat low bone density [39,49,50]. Other hormone replacement therapies for the endocrine consequences of AN, such as growth or pubertal delay, appeared in the 2000s in association with endocrine-paediatric management in the context of a multidisciplinary approach [151,157]. Two recent narrative update reviews on endocrine mechanisms and repercussions in AN report on available treatments and innovative therapeutic strategies in endocrinology [189,190]. Physicians' interest in refeeding modalities [191], relating to optimal daily calorie requirements [42], the risk of over-feeding and that of under-feeding, high calorie oral supplementation [116], vitamin supplementation and functional digestive disorder medications [139], is fairly recent and corresponds to the development of specific nutritional and dietary expertise, alongside the publication of international guidelines [2,3]. Finally, in the last 20 years, we have observed a paradigm shift concerning enteral tube feeding, considered initially as unethical and coercive, and now viewed as efficient, safe and well tolerated [53]. Promising somatic medications for new targets have developed recently, with micronutrients such as zinc, polyunsaturated fatty acids (PUFAs) used in refeeding, physiological oestrogen replacement or teriparatide to treat bone health, leptin to restore gonadic function, reproductive hormone for hormonal deficits and neuropsychic functions, and GH for severe growth retardation among prepubertal adolescents.

#### *4.2. Psychotropic Drugs*

Since the 1960s nearly all classes of psychotropic drugs have been tried in AN [45]. Psychotropic drugs were initially compared to placebo without psychotherapy, with very poor results, but the most recent studies now use psychotropic drugs as additional treatment to global treatment approaches including psychotherapy (individual and/or family therapy). This type of use stems from an international consensus, as attested by the NICE guidelines (the most recent international guidelines published) stipulating that medication cannot be considered as the sole treatment for Anorexia Nervosa (see Table 1).

Despite the evolution and improvement in study design, evidence-based data is scarce and possibly even less robust than it was in the WFSBP in 2011 [6] since atypical antipsychotics seem not to exhibit the efficacy they initially appeared to have.

The RCTs generally evidence negative results concerning weight, and contradictory results concerning eating disorder symptoms and psychiatric dimensions. This situation can be explained by discrepancies in study methodologies [45,48], including many limitations such as relatively short durations of treatment (weeks, or two to three months mainly) and the small sample sizes (fewer than 100 in RCTs, and generally fewer than 30). That said, one of the most important limitations to efficacy in AN is non-compliance with medication treatments: patients frequently do not take their medication.

This situation of a poor evidence base in the existing literature is discouraging [48]. It has led, in the more recent reviews published, to giving prominence to emergent or promising interventions and recommendations for research [24,40,45,51].

In this situation, if we refer only to the evidence-based elements, psychotropic drugs should not be prescribed in AN. However, the reality in the field is quite the reverse. All psychotropic drugs are widely prescribed to both adults and adolescents with AN, and this seems to be increasing [7]. In addition, more than half of the patients have more than one psychotropic drug prescribed simultaneously. Also, dangerous drugs such as bupropion or tricyclics, are still prescribed despite the conclusions of the literature [7]. The same observation has been made in other studies on adults and adolescents in different countries [7,8,192]. These papers highlight the disparity between research conclusions and clinical practice in the treatment of AN. This underlines the importance of consensual guidelines incorporating evidence-based data. We also need expert consensus and training for practitioners on this question, especially for ED specialists, who mostly base their practice on their experience and who seem more resistant to using guidelines than general practitioners [7]. Our "experience" is the result of a complex combination of elements. It is based on core beliefs derived from different sources, including, of course, evidence-based data. Evidence can be biased in favour

of drug efficacy by the publication of positive results only [7]. In addition, some studies (or their coordinators) are sponsored by the industry and their results can be biased by conflicts of interest. This last point leads the latest NICE guidelines to exclude studies sponsored by the industry from the studies analysed to establish the evidence base retained for the guidelines [4]. In addition, our experience is also impacted by psychiatric theoretical orientations, [7] or good or bad experiences with one or several patients.

## *4.3. Use of Psychotropic Medication in AN*

Once we have discussed these elements, the central question for this review is still unresolved: how should we use medication in AN?

As recommended by most guidelines, we need to evaluate the global situation of the patient, including psychiatric, medical, nutritional and social aspects, to propose the best-suited treatment. Medication should be prescribed on the basis of the clinical evaluation. This evaluation should include the patient's opinion about the treatment [4,24].

Concerning psychotropic medication, comorbidities (including mood disorders, anxiety disorders, obsessive disorders) should be evaluated and treated [18,45] if they are real disorders and not "artefacts" that will improve with re-feeding [193]. These disorders are usually treated by teams on the assumption that treating anxiety and depressive disorders facilitates AN treatment, or at least improves their quality of life [45]. These disorders are usually treated according to the specific recommendations for each, and taking into account the particular risks of side effects resulting from low weight and metabolic disturbances (possible depletion in potassium, phosphorus, magnesium, and zinc), and somatic complications concerning hematologic, renal, hepatic and heart functions that can interact with medication. Antidepressants and atypical antipsychotics are the most frequently evaluated drugs and the most widely used in practice.

Chlorpromazine is no longer used for AN because of its severe adverse effects, including seizures; low doses of haloperidol could be effective as an adjunct treatment for patients with severe, treatment-resistant conditions with marked self-image alteration [48]. Atypical antipsychotics can sometimes be useful to help patients with AN, in case of high levels of anxiety [75] or in adults with a long illness duration [76], but practitioners should be aware that they prolong the QT segment and can be dangerous in case of malnutrition by leading to cardiac death associated with ventricular arrhythmia. Their prescription should be monitored for safety, including regular electrocardiogram and ionogram assessments. In addition, because of the risk of metabolic syndrome, such as hyperglycaemia, dyslipidaemia or HTA, the usual guidelines for the monitoring of antipsychotics should be applied. Benzodiazepines are also widely used and their use is questionable, especially among young people, as they induce dependency and can lead to abuse or misuse [7].

A psychotropic drug can be useful to manage eating disorders or psychiatric symptoms. Himmerich et al. [194] in a recent survey among AN patients and carers and reported that people with anorexia nervosa want medication to help with anxiety and sleep problems [24]. However, the specific treatment of AN also requires re-feeding, and/or cessation of binging and purging, specific psychotherapy (individual or family therapy) and work on the social impact of the illness. Medications, and especially antidepressants, are less effective among those that are acutely ill and underweight [18].

Outside situations of emergency (suicidal ideas or risk, very severe anxiety or OCD), following the discussion with the patient, it is possible to wait a few weeks in order to later reassess the need for specific treatments for possible comorbidities [195]. In order to diagnose whether psychiatric symptoms are linked to a comorbid diagnosis or to malnutrition, it is useful to investigate the personal and family history of comorbid disorders and the chronology of appearance of AN and psychiatric symptoms. Diagnosis should be based on symptoms, their evolution, the chronology of onset and individual and familial history of comorbidities [18].

A general discussion with the patient (and the parents for a minor) is necessary to review side effects, risks, benefits, and alternatives. One important point to make is that psychotropic drugs do not induce weight gain in AN but can alleviate negative psychological symptoms. This will be important for the therapeutic alliance. Practitioners should always have in mind that to be efficient, a medication needs be taken, and AN patients are often not observant, especially for antipsychotics [37]. In addition, to be efficient, a medication needs be absorbed. When patients purge after they have taken their pills or have swallowing phobia or food orality disturbances, they cannot be efficient. Both points should be discussed with the patients.

As in other reviews, we will conclude on perspectives for future research on the topic of medication in AN.
