Catatonia is a severe form of psychomotor disturbance with a heterogenous presentation. It affects approximately 10% of acute psychiatric inpatients. According to the fifth edition of DSM-5 the diagnosis of catatonia can be made when three or more symptoms from the twelve following are present : catalepsy, waxy flexibility, stupor, agitation, mutism, negativism, posturing, mannerisms, stereotypies, grimacing, echolalia, echopraxia. It can occur in various psychiatric diseases, including mood disorders or schizophrenia, but also in various non-psychiatric disorders \[metabolic disturbances, viral infections (including HIV), typhoid fever, heat stroke, and autoimmune disease\]. Benzodiazepines, especially LORAZEPAM, are the most common initial treatment, with a remission rate of approximately 70-80 %, regardless of the cause or the clinical manifestations. This first line treatment is titrated gradually according to the therapeutic response over a few days up to 20-25 mg per day. Electroconvulsive therapy (ECT) is initiated on patients with catatonia who do not respond to benzodiazepines. Interestingly, pharmacogenetic variants can alter the metabolism of lorazepam (e.g., the UGT2B15 \* 2 allele slows it down). The main objective of this study is to assess the link between clinical response to lorazepam, residual plasma concentrations of lorazepam after 72 hours of fixed dosage, and the existence of genetic polymorphisms modifying the metabolism of lorazepam. Our hypothesis is that non-responding patients have lowered blood concentrations of lorazepam associated to a genetic profile of rapid metabolism. Evaluating the predictive factors of the response to treatment would allow early and precise identification of non-responder patients in order to adapt their first-line treatment.