BACKGROUND Depression is common in people with schizophrenia and is associated with substantial morbidity explaining also the considerable attention and recognition of this entity as suggested by the inclusion of the post-psychotic depression in DSM IV and ICD 10. The prevalence of this disorder varies according to the type of approach used (range between 7% to 75%). Prescription of antidepressants plus antipsychotic treatment is frequent in clinical practice (11 to 43%). BACKGROUND Pharmacokinetic and metabolic interactions have been identified. The cytochrome P450 has been identified as being implicated in the metabolism of most psychotropics, mainly through the CYP1A2, CYP2C19, CYP2D6, CYP3A4 isoenzymes. Tricyclic antidepressants are likely to increase chlorpromazine plasma levels. Similarly, antipsychotics such as perphenazine, chlorpromazine or haloperidol can increase antidepressant plasma levels, through the inhibition of CYP 450 isoenzymes (CYP2D6). Most of the Specific Serotonin Recapture Inhibitors (SSRIs) are likely to inhibit one or several CYP450 isoenzymes. The inhibition is moderate to marked for CYP1A2 (fluvoxamine and fluoxetine), CYP2C19 (fluoxetine, fluvoxamine and sertraline), CYP2D6 (paroxetine, fluoxetine and sertraline), and CYP3A4 (fluvoxamine, fluoxetine and sertraline). In the US, one-fourth of psychiatrists report the use of depression-rating scales in schizophrenic patients. Non specific scales (Hamilton Depression Rating Scale or Beck Depression Inventory) are the most commonly used in spite of the fact that these scales do not allow the distinction of depressive from negative symptoms in schizophrenic patients. RESULTS Due to these limitations, more specific assessment tools for depressive symptoms in schizophrenia are required. Two specific scales for assessing depressive symptoms in schizophrenic patients have been constructed and validated. The Calgary Depression Scale (CDS) is a nine item scale, each item scored from 0 to 3. This scale was derived from the HDRS and the Present State Examination. Factor analysis showed that the CDS is unidimensional, has high internal consistency, and significant strong correlation with scores on the HDRS, Beck and BPRS depression scales. The CDS has been validated in different languages (Brazilian, Danish, French...). It has been shown that there is no overlap between negative or extrapyramidal and depressive symptoms assessed by the PDS in schizophrenic patients. The Psychotic Depression Scale (PDS) is a 32 item scale derived from the HDRS, PANSS, CPRS and AMDP, each item being rated from 0 to 7. A principal component analysis of the PDS items using a Varimax rotation disclosed 8 orthogonal components that account for 71% of the variance. These components involved the following dimensions : depressive mood, inhibition, vegetative signs, paranoid signs, strangeness of thought, inverse vegetative signs, guilt feelings and cognitive signs. The analysis revealed that the 'depressive mood' factor of the PDS was correlated with the 'depressive' factor and was slightly correlated with the cognitive factor of the PANSS. This first factor was not correlated with either the "negative" factor of the PANSS, or the Positive or Excitement factor of the PANSS. Hence, this PDS, factor distinguished depressive signs from negative symptoms. Due to their metrologic properties, specific scales should be preferred. However, only one open trial (of an antipsychotic) and two double blind controlled trials (one comparison of 2 antipsychotics and one comparison of an cholinesterase inhibitor versus placebo) have been published using the CDS. Likewise, only one double blind controlled trial using the PDS (comparison of 2 antipsychotics) has been published. No study of the effect of antidepressants in depressed schizophrenic patients has been published, using either the CDS or the PDS assessment criteria. CONCLUSIONS These specific scales are rarely used in clinical practice. Only about 1% of the US psychiatrists reported the use of the Calgary Depression Scale. Several open clinical trials have assessed the efficacy of antidepressant agents added to antipsychotic in patients with schizophrenia. They have produced inconsistent results but have suffered from methodological limits (short duration of the studies, non homogeneous inclusion criteria, heterogeneous assessment methods...). Due to the lack of a reference drug, double blind placebo-controlled trial are necessary. A recent meta-analysis has been performed on results of trials that have investigated the clinical efficacy of antidepressant medication (either tricyclics, SSRIs or others) in the treatment of depression in schizophrenic patients. In a subset of 5 trials (209 patients), the proportion improved in the antidepressant group was 26% (95% CI 10 to 42) higher than in the placebo group. The estimated number needed to treat was 4. In a subgroup of 6 studies (267 patients), the standardized mean difference on the HDRS score was -0.27 (95% CI - 0.7 to -0.2). There was no evidence that antidepressant treatment induced a deterioration of psychotic symptoms in these trials. CONCLUSIONS The results provide weak evidence for the efficacy of antidepressants in patients with schizophrenia and depression. Today, the only SSRI tested in the treatment of depression in schizophrenic patients is sertraline. One study led to positive results. Since the meta-analysis, one additional study has been performed comparing sertraline to placebo. No difference between the 2 treatment groups was demonstrated but the power of the trial was rather low. Further research is required to determine the best approach towards treating depression in patients with schizophrenia, with clinical trials performed for longer periods, using appropriate assessment criteria such as depressive symptoms and quality of life.