Meta-analysis of regional white matter volume in bipolar disorder with replication in an independent sample

Background and Aims Converging evidence from a number of neuroimaging techniques suggests that bipolar disorder (BD) is associated with white matter (WM) abnormalities, however a meta-analysis of WM voxel based morphometry (VBM) studies has not been conducted. Obtaining statistical maps from studies (T-maps) remedies the limitation of meta-analyses using published coordinates only and substantially increases accuracy. Acquiring original MRI datasets increases accuracy further and allows a comprehensive analysis of clinical variables. Methods Following a literature search of VBM studies of BD, authors were contacted for additional data. In addition to published coordinate data from 5 studies, we obtained 8 T-Maps and 5 original MRI datasets. A total of 18 studies of WM volume were meta-analysed using the software ES-SDM. The result of the meta-analysis was validated in an independent dataset (26 patients, 23 controls), and the effect of clinical variables was examined using all original MRI data (184 patients). Results The meta-analysis revealed decreased regional WM volume in the posterior corpus callosum extending to white matter tracts and adjacent to the posterior cingulate cortex. This region was significantly reduced in volume in patients from the independent dataset (p=0.0006) and remained significant when controlling for lateral ventricle volume. Using the original MRI data we found no association in this region with age of onset, duration of illness, lithium use, antipsychotic use, history of psychosis, bipolar disorder type or age. Conclusions In summary we identified a robust regional WM volume reduction in patients with BD and used a novel methodology to validate the findings in an independent dataset.status: publishe