Whole brain tractography imaging shows decreased white matter integrity in patients with bipolar I disorder, compared with individuals without the disorder, a study published online in JAMA Psychiatry shows.
This effect was even more pronounced in bipolar patients with a history of psychotic features, defined as the patient having had at least one manic or depressive episode with delusions or hallucinations, wrote Samuel Sarrazin and his colleagues.
The investigators compared generalized fractional anisotropy (GFA) values in 118 bipolar I patients recruited from multiple university-affiliated centers in France, Germany, and the United States, with a control group of 86 participants who had no personal or family history of Axis I mood disorders, schizophrenia, or schizoaffective disorder. A lower GFA value suggests loss of integrity or coherence of white matter, the authors said.
At all of the sites, the Montgomery-Åsberg Depression Rating Scale or the Hamilton Depression Rating Scale were administered, in addition to the Young Mania Rating Scale, and the National Adult Reading Test, reported Mr. Sarrazin, who is affiliated with several French medical institutions, including the Assistance Publique–Hôpitaux de Paris.
The mean age of disease onset among the patients with bipolar I was about 20.8 years, and the mean age at MRI was about 36.3 years. For the healthy controls, the mean age at MRI was 37.2 years (JAMA Psychiatry 2014 Feb. 12 [doi:10.1001/jamapsychiatry.2013.4513]).
In a linear mixed-model analysis comparing the two groups, patients with bipolar I disorder had an average GFA value of 0.101 in the body of the corpus callosum, 0.113 in the splenium, and 0.079 in the anterior segment of the left hemisphere, compared with values of 0.102, 0.115, and 0.081, respectively, in the control group. Results remained significant when adjusting for false discovery rate (P = .03), and were consistent when researchers removed patients from the sample who had confounding factors such as elevated or mixed symptoms, the taking of lithium, and the existence of past alcohol abuse.
Bipolar I patients with a history of psychotic features had even lower GFA values than did bipolar patients with no psychotic history, with a mean GFA value of 0.100 in the body of the corpus callosum. This difference remained significant after adjusting for false discovery rate (P = .03).
"These results highlight the role of interhemispheric disconnectivity" in bipolar I disorder and suggest that bipolar I disorder with psychotic features "could be a relevant subtype of bipolar disorder with specific pathophysiological features," the investigators wrote. Additional large, multicenter studies are needed to compare bipolar I disorder with other psychotic disorders, and to further study such neuroimaging biomarkers, they added.
In an accompanying editorial, Dr. Kathryn R. Cullen and Dr. Kelvin O. Lim said that the latest findings advance the diffusion imaging literature in that the sample was large enough to address two important questions: "where [white matter] deficits are most consistent and how abnormalities vary across subtypes of [bipolar disorder]."
Furthermore, they said, the findings "conclusively affirm prior reports suggesting impaired [fractional anisotropy] in [white matter] tracts ... using state-of-the-art methods. Exciting contributions are the documentation of a more severe biological abnormality in the subgroup of patients with psychosis and additional evidence supporting an interhemispheric disconnectivity theory in [bipolar disorder]," wrote Dr. Cullen and Dr. Lim, both of the University of Minnesota, Minneapolis (JAMA Psychiatry 2014 Feb. 12 [doi:10.1001/jamapsychiatry.2013.4638]).
Mr. Sarrazin and his colleagues cited several limitations to their study. First, they "did not explore the interrater and intersite reliability of the scales used in the study." Second, they did not include a "phantom procedure" to check the intercenter quality of the acquisitions. Third, the authors cannot exclude a possible effect of past medication use on the results. Fourth, given the large number of tracts to assess, the authors did not exploit other metrics from the orientation distribution function. Last, mean GFA values were calculated along each tract to perform comparisons, and localized decreases in GFA values might have gone undetected.
Neither Mr. Sarrazin nor his colleagues reported conflicts of interest. The study was funded by Alliance Nationale pour les Sciences de la Vie et de la Santé (aviesan), the Agence Nationale pour la Recherche, the Deutsche Forschungsgemeinschaft, the National Institute of Mental Health, and the Agence Régionale de Santé Ile-de-France.