Phenotype homogeneity and heritability are important conditions for identifying the genetic basis of bipolar I disorder (BPI) in association studies. Our objective was to study the heritability of mood-incongruent psychosis (MIP) in BPI in a sample of 504 families ascertained through BPI probands (294 females; 210 males) recruited from consecutive hospital admissions.

There were 402 families with a psychotic proband and 275 families with a proband with MIP. All probands were directly interviewed as well as 79.55% first-degree and 22.59% second-degree relatives. The narrow and the broad sense heritability (h2) of MIP and the effect of sex and age were estimated using S.A.G.E.v.6.01-software (2009).

There was no sex difference for the psychosis prevalence in probands but MIP was two times more frequent in females than in males. In families with MIP probands the narrow-sense h2 for MIP was 0.14 (SE = 0.02, P = 0.002) and the broad-sense h2 was 0.20 (SE = 0.014, P = 0.0000). Significant but lower heritabilities were found in families with a psychotic proband (narrow-sense h2 = 0.12; broad-sense h2 = 0.13). In the total sample the narrow-sense h2 was 0.06 (P < 0.005) and the broad-sense h2 was 0.10 (P < 0.00001). The female sex was more prone to incongruency (χ2 = 33.32, P = 0.0000).

The heritability of MIP was significant but not high in families ascertained through BPI probands regardless of familial psychopathology. These finding is in line with GWAS-studies showing that the polygenic score fails to differentiate psychotic BPI from non-psychotic BPI. Is therefore incongruent psychosis a useful dimension for association studies?