Depression and anorexia nervosa are often co-morbid. According to a recent meta-analysis [5], there is no evidence for the effectiveness of antidepressants among patients with AN in terms of depressive symptoms regardless of the type of prescribed treatment (tricyclics, Selective Serotonin Reuptake Inhibitors, other antidepressants as monoamine oxydase inhibitors, venlafaxine). This meta-analysis does not mention ketamine nor esketamine because of the rarity of cases. The authors conclude that these psychotropic drugs should not be prescribed in AN.

An explanation for this problematic antidepressants’ ineffectiveness could be linked to malnutrition [9]. Several authors hypothesised that undernutrition and its consequences lead to depression and to treatment-resistance issues.

Abnormalities in central nervous system monoamine metabolisms were highlighted [10]. Some hypotheses are related to the lack of dietary tryptophan an essential amino acid, which is the precursor of serotonin (5-HT). Some studies indeed found a reduced level of 5-HT metabolite in the cerebrospinal fluid of animals and healthy humans under a restricted diet [9,10]. A few years later, in a simultaneous exploration of depression, nutritional status and the peripheral metabolism of tryptophan and serotonin, before and after re-feeding and weight gain, some authors gave further evidence that tryptophan availability could restore serotonin neurotransmission and lead to depression improvement after re-feeding [11]. Impaired serotonin pathway could also be triggered by vitamins dietary deficiency such as B6 or B12 vitamins, both being involved in 5-HT synthesis [9].

Another possible explanation is the lack of zinc, a mineral obtained from dietary sources (especially red meat, poultry, fish, and dairy products). Zinc is an allosteric modulator that physiologically blocks post-synaptic glutamatergic NMDA receptor functioning [9] and zinc levels are known to be higher in limbic structures as in the amygdala, hippocampus and across the cortex [12]. Low zinc levels are associated with treatment resistance due to malnutrition in AN [8] and were also linked to the risk of major depression [12,13,14]. At a cellular level, the lack of zinc leads to modified glutamatergic neurotransmission via increased NMDA receptor activity in several ways, including (1) decreased stimulation of GABA interneurons -which inhibits glutamate release from presynaptic neuron – leading to increased glutamate level in the synaptic cleft; (2) decreased inhibition of NMDAr, leading to increased levels of Calcium into post-synaptic neuron; and (3) decreased inhibition of metabotropic glutamate receptor, leading to an increased release of Calcium. Those effects, all together, result in an upregulation of NMDAr, that leads to an increased Calcium influx into the postsynaptic neurons, which can cause excitotoxicity, and therefore impair synaptogenesis and synaptic plasticity [13]. Zinc supplementation alone did not lead to AN and/or MDD remission, but adjunctive supplementation could be a helpful therapeutic strategy for MDD [5,12].

By inhibiting NMDA receptor, ketamine and esketamine could lead to “normal” state and avoid excitotoxicity. Ketamine and its enantiomer S-ketamine are NMDAr antagonists and have better biodisponibility when administered intravenously (100% biodisponibility) or by nasal spray (45% biodisponibility) compare to sublingual (30%) or oral (20%) [15]. Ketamine is now an acknowledged treatment for unipolar and bipolar depression and esketamine recently obtained FDA approval for treatment-resistant unipolar depression [15].

Its rapid antidepressant action contrasts with usual antidepressant treatments. Moreover, its specific action on suicidal ideation makes ketamine a particularly useful treatment to prevent suicidal behaviours in patients with anorexia nervosa who have one of the highest suicidal risk of death (Chesney et al., 2014).

A review assessed potential mechanisms of this rapid pharmacological action: AMPAr activation, NMDAr inhibition-mediated mechanisms, inhibition of NMDARs expressed on GABAergic interneurons (disinhibition hypothesis), inhibition of spontaneous NMDARmediated transmission, direct inhibition of extra-synaptic NMDAR [16]. Moreover, ketamine administration induces a rapid increase in total BDNF protein levels, within 30 minutes of administration. Thanks to its direct glutamatergic effect, ketamine might be efficient independently of nutritional state. Interestingly, there is convergent evidence that patients with anorexia nervosa have an excess of the Met-allele, the low activity BDNF genetic variant (Ribasés et al., 2005 ; Gratacòs et al., 2007; Mercader et al., 2007 ; Ceccarini et al., 2019), and BDNF serum concentrations are significantly higher in recovered AN participants compared to acute AN patients (Zwipp et al., 2014). Furthermore, intravenous or intranasal galenic allow better compliance avoiding lack of efficacy due to purgative behaviours.