La stratégie dans le choix d'un antipsychotique doit souvent prendre en compte la tolérance hépatique du fait d'une incidence non négligeable de troubles préexistant chez les patients psychotiques (facteurs de risque parfois retrouvés dans cette population tels que alcoolisme, pratique addictive, polymédication incluant des principes actifs à potentiel hépatotoxique). La prescription des antipsychotiques atypiques disponibles en France (amisulpride, clozapine, olanzapine, rispéridone) s'accompagne dans la très grande majorité des cas d'une bonne tolérance hépatique. L'incidence des troubles hépatiques varie selon les agents mais reste toujours très peu fréquente ou rare. Toutefois, certains cas ont été publiés dont le diagnostic et l'imputabilité sont rendus difficiles par un délai d'apparition très variable, de un jour à plusieurs mois, et par une fréquente polymédication. Les principaux cas ont été rapportés avec la clozapine, l'olanzapine et la rispéridone. Les troubles hépatiques ont généralement été réversibles à l'arrêt du traitement sauf pour un patient de 39 ans traité par clozapine. Les troubles hépatiques étaient parfois associés à d'autres désordres de type neurologique, métabolique ou généraux. Du fait du risque majeur parfois irréversible pouvant être associé à la réintroduction, l'antipsychotique a été arrêté définitivement dans de très nombreux cas. Chez certains patients pour lesquels la perturbation était limitée (élévation isolée des enzymes hépatiques), l'effet est apparu réversible malgré le maintien de l'antipsychotique. Certains facteurs de risque ont pu être parfois retrouvés, parmi lesquels l'utilisation de forte posologie ou un surdosage, l'âge (gériatrie ou pédopsychiatrie), l'obésité, l'association à des principes actifs ou des substances addictives responsables de troubles hépatiques. La mise en place d'études multicentriques présenterait l'intérêt de comparer les molécules entre elles en termes de fréquence de survenue d'une atteinte hépatique et de tolérance à long terme et permettrait de faire ressortir d'éventuels facteurs de risque.

The strategy in the choice of antipsychotic agent must take into account the hepatic tolerance according to non-negligible incidence of liver disorders among psychiatric population (presence of risk factors like alcoholism, drugs of abuse intake, polymedication including potentially hepatotoxic drugs…). More than 1 000 drugs have been listed as being responsible of hepatic side effects ; 16 % of these agents were neuropsychiatric drugs. Antidepressive drugs (tricyclic agents or SSRI), mood stabilizing agents and neuroleptic drugs have been implicated in biological or/and clinical hepatotoxicity. For these reasons, some psychotropic agents have been withdrawn of the pharmaceutical market like alpidem or medifoxamine. Atrium*, sometimes used to correct tremor induced by neuroleptic drugs, has been withdrawn recently, as well. Isolated elevations of hepatic enzymes occur frequently with phenothiazines drugs (frequency evaluated to 20 %) but also with other classes of neuroleptic agents, as well. On the contrary, clinical hepatitis have been more rarely described with neuroleptic drugs like phenothiazine agents (0,1-1 %) or with haloperidol (0,002 %). The definition of hepatotoxicity is based on biological parameters (elevation of alkaline phosphatase enzyme, SGPT, SGOT and GGT) or on clinical abnormalities (hepatitis, jaundice…). Clinical hepatitis could be either cytolytic or cholestatic. Clinical diagnosis and the research of its origin may include many investigations like abdominal ultrasonogram and percutaneous liver biopsy. The present article describes the cases of hepatic disorders reported with AAD (Atypical Antipsychotic Drugs), which are available in France (amisulpride, clozapine, olanzapine, risperidone). This new pharmacological class of antipsychotic drugs has showed great interest to improve negative symptoms of schizophrenia and to reduce disabling side effects like dystonia. According to the bibliographic data available, the following points and information must be clinically taken into account. Frequency of hepatic troubles : according to the bibliographic data, AAD appeared generally well tolerated in most cases. The frequency of hepatic troubles remains in general very low or rare. The cases published were observed with clozapine, olanzapine and risperidone. Nevertheless, some authors have observed higher frequency of hepatic enzymes elevation with some AAD. In an investigation comparing hepatic tolerance of clozapine (n = 167) versus haloperidol (n = 71), 37,3 % of clozapine treated patients showed a relevant SGPT increase versus 16,6 % with haloperidol. Nature of the hepatic troubles : among the clinical observations, asymptomatic biological disorders of the hepatic function are generally described but cytolytic or cholestatic hepatitis were reported, as well. Symptomatic hepatic dysfunctions were, sometimes, associated with other disorders like convulsions, pneumonia or malignant syndrome. Thus, hepatic check-up may be relevant in case of significant side-effect outcome. Delay time before the hepatic episode : hepatic injuries generally occurred within the first weeks of treatment but this delay highly varied in the literature from 1 to 8 weeks, 12 days to 5 months, 1day to 17 months for clozapine, olanzapine and risperidone, respectively. These delay times are very similar to those observed with other psychotropic drugs. Reversibility of the hepatic troubles and rechallenge of the responsible agent : all cases were reversible after the AAD withdrawal except with one patient (39 years old) treated by clozapine (350 mg/day) who developed a fulminant and irreversible hepatitis after 8 weeks of monotherapy. In most cases, the AAD was withdrawn after the hepatic episode according to the significant risk of irreversible alteration. Nevertheless, normalization of hepatic enzymes has been described despite AAD maintenance at the same dosage or after dosage reduction. Rechallenge of clozapine after a first episode was performed for three patients, only one redeveloped a new hepatic disorder. According to different authors, special care is required if maintenance or rechallenge of the agent is indispensable after a first episode of isolated hepatic enzyme elevation (i.e resistance or intolerance to other treatments). In this case, biological and clinical supervision has to be carefully scheduled, which demands a satisfactory compliance from the patient. On the contrary, in case of clinical hepatotoxicity, rechallenge or maintenance is absolutely inadvisable. Mechanism of the hepatic troubles : precise mechanisms of the hepatotoxicity remain unclear. Contrary to phenothiazine drugs, no information is available on the respective rule of the agents and their metabolites. Hypersensitivity syndrome or eosinophilia has been reported, suggesting a possible immuno-allergic mechanism. Presence of risk factors : risk factors have been retrieved, in some observations, like high daily dosage, high plasmatic concentration, age, alcoholism, obesity or antecedent of hepatic disorders like Gilbert syndrome. Special care is advisable with these patients. As hepatotoxicity has been observed after surdosage (or suicide attempt), a hepatic check-up has to be performed in these clinical situations. Co-medication with hepatotoxic drugs may increase the risk as it has been suggested. In many observations, co-medication made difficult the incrimination of the AAD in the hepatic disorders outcome. Monotherapy has the great advantage to make easier the withdrawal of the responsible agent and its substitution. As drugs of abuse like cocaine or ecstasy are notoriously responsible of hepatotoxicity, they represent a probable factor of risk. Moreover, their detection is fundamental during the clinical investigation. Conclusion – Diagnosis of toxic hepatitis is mainly based on the chronology between agent introduction and hepatic disorder onset but other causes must be excluded. Bibliographic data analysis greatly contributes to confirm toxic hepatitis diagnosis. Nevertheless, this article emphasized the limits of bibliographic review to compare drugs towards tolerance. Most of the bibliographic data were case-reports for which it was sometimes difficult to provide absolute evidence of the responsibility of the agent. Moreover, spontaneous notification to health national administration is rarely systematic, in particular with isolated elevation of hepatic enzymes, and even more rarely published in international reviews. Nevertheless, according to the present data available in the literature, systematic and regular hepatic survey does not seem necessary in absence of risk factors. As for other side effects, which may occur more or less rapidly, great advantages may be obtained from psycho-education programs associating the patients in order to detect the first symptoms. Because little long-term hepatic follow-up comparing AAD is available, controlled studies should be carried out to precise the frequency and the risk factors (covariables) to prevent hepatitis outcome.