The response to antipsychotic drugs could be one attempt to develop a candidate phenotype of schizophrenia and deconstruct the illness for genetic analysis. This review discusses six studies on the association between HLA and the response to antipsychotic drugs. Three studies implicate an association between class I antigens. However, there are considerable limitations due to the small sample size, the different genetic backgrounds of the study subjects, varying diagnostic criteria, lack of standard criteria for drug response or refractoriness, inappropriate statistical methodology and varying HLA typing methods. Only for HLA-A1 was an association with response to standard neuroleptic treatment found repeatedly, indicating that the presence of HLA-A1 could predict a poorer outcome with conventional neuroleptic treatment. An association between HLA specificity and antipsychotic drug response may represent linkage disequilibrium with genes predisposing to schizophrenia on chromosome 6. Further studies are warranted with larger sample sizes in ethnically homogeneous populations, and standard criteria should be used for characterising the subjects.