Neurocognitive endophenotypes (endophenocognitypes) from studies of relatives of bipolar disorder subjects: a systematic review.

BACKGROUND: There is growing interest to research neurocognition as a putative endophenotype for subjects with bipolar disorders (BD). The authors sought to review the available literature focused on relatives of subjects with bipolar disorder (BD-Rels) and identify suitable cognitive candidates to endophenotypes or endophenocognitypes.
METHOD: A systematic review was conducted in Medline, EMBASE and PsycINFO databases (1980-July 2007), supplemented with a manual search of reference lists.
RESULTS: Twenty-three cross-sectional papers of discordant twins (4 studies), genetic high-risk subjects (7), and different BD-Rel groups (12) met the inclusion criteria and evaluated 532 BD-Rels. Impairments on the broad domain of verbal learning/memory were found in 6 out of 11 studies (54%), as well as in 3 of 9 reports (33%) of working memory. Moreover, BD-Rels showed deficits in visual-spatial learning and memory (1/6 reports; 17%), alternating attention (1/8; 12.5%), psychomotor speed (2/10; 20%), and abstraction/cognitive flexibility, sustained attention and selective attention (2/8 each; 25%). Scores of general intelligence were lower than those of controls in 2/16 (12.5%) reports, but fell well within the average range in all studies. No study that assessed immediate memory or verbal fluency (6 each) reported impairments in BD-Rels. Finally, language, social cognition, and motor and planning skills are neglected areas of research.
CONCLUSIONS: Overall, the neurocognitive profile in BD-Rels is still unclear, and the evidence in support of the presence of cognitive deficits seems quite sparse. Verbal learning/memory and verbal working memory seem to be the most suitable endophenocognitypes for BD. Conversely, healthy family members would have an intact performance on immediate memory, verbal fluency, and probably on general intelligence. The possibility that BD-Rels show less cognitive efficiency compared to healthy controls also on other functions must be addressed by future studies with larger samples, comprehensive neuropsychological assessments, and, ideally, longitudinal designs.