Predictive biomarkers are tools that identify a subpopulation of patients who are most likely to respond to a given therapy. In order to identify them a strict methodology is necessary (RCT's studies). In consideration of its cost in economic and medical terms, predictive biomarkers would be useful to distinguish clozapine-resistant patients before its administration.

The evidence concerning genetic biomarkers was reviewed with the aim of assessing whether there is enough evidence to claim for predictive biomarkers useful in practice. Secondary aims were the assessment of the evidence concerning genetic prognostic biomarkers and predictors of side effects in clozapine-treated schizophrenic patients.

One hundred and twenty-eight studies, searched on the Pubmed database or referenced in other studies, were included in this review. Sixty-five papers were related to clozapine efficacy and explored 167 genetic variants.

Fifty-four variants were supported as prognostic biomarkers, three were successfully replicated: rs6280, rs6314 and rs4680; 49 papers were related to clozapine weight gain and explored 216 different genetic variants. Forty-five of which were positively related to weight gain during clozapine treatment. Among these 45 variants, only two, Rs3813929 and Rs779039, were successfully replicated.

Fourteen studies explored 111 genetic variants potentially correlated to Clozapine-induced agranulocytosis. Thirty-four variants were found to be associated with agranulocytosis. Five variants had positive results, successfully replicated. In particular, HLA B38.

To date there is no evidence to support a modification of clinical practice towards predictive medicine. The research could ideally progress with RCTs involving the prognostic factors found in association studies.

The authors have not supplied their declaration of competing interest.