Reduced left hemispheric white matter volume in twins with bipolar I disorder

Abstract

Background

Although the heritability of bipolar I disorder (BPI) is high, few magnetic resonance imaging (MRI) studies of siblings of bipolar patients exist. We performed MRI brain scans on a nationwide sample of twins with BPI, as well as on their co-twins and a demographically balanced sample of control twin subjects, to detect any structural alterations related to the disorder and to the increased genetic risk.

Methods

The National Hospital Discharge Register, National Population Register, and Finnish Twin Cohorts were used to identify bipolar twins. Structured diagnostic interviews and MRI scans were obtained for 24 twins with BPI, 15 healthy co-twins, and 27 control twin subjects.

Results

Patients and co-twins showed a significant decrease in left hemispheric white matter volume. The disparity in patients was −16.1 cm³ (95% confidence interval [CI] −26.6, −5.6) and in co-twins −11.3 cm³ (95% CI −22.1, −0.4) compared with control twin subjects. No gray matter decrease was seen in patients or co-twins.

Conclusions

The results of this first large-scale MRI study of twins with BPI, their co-twins, and appropriate control twin subjects, suggest that alterations of the left hemisphere white matter in BPI may reflect genetic factors predisposing to the disorder.

Introduction

The improvement of magnetic resonance imaging (MRI) techniques offers intriguing possibilities for studying the relationship between psychiatric disorders and neuropathology. Recent reviews Bearden et al 2001, Beyer and Krishnan 2002, Pearlson 1999, Strakowski et al 2002 of structural brain changes in bipolar I disorder (BPI) report a large variety of findings, but their consistency seems to be vague, and the etiology remains unresolved. Possible explanations for the discrepant nature of the findings are discussed widely in the reviews; however, all four reviews point out the abnormalities found in the frontal lobe and the plausible implication that frontal lobe, limbic, and basal ganglia circuits play a role in the pathology of BPI. One volumetric MRI study of patients with BPI did indeed report a trend of smaller frontal area (Coffman et al 1990), but two other studies found no significant frontal differences compared with control subjects Strakowski et al 1993, Zipursky et al 1997. However, the above reviews argue that alterations might be subregion- or subgroup-specific. Interestingly, two studies reported smaller left subgenual cingulate areas in a familial type of BPI Drevets et al 1998, Hirayasu et al 1999, but one study did not (Brambilla et al 2002). Decreased prefrontal volume in BPI patients, which was associated with impaired performance in the Continuous Performance Test, was found by Sax et al (1999).

The theory that neuroanatomic circuitry is implicated in the pathology of bipolar disorder would imply the involvement of the temporal lobe, particularly the hippocampus and amygdala. Volumetric studies report both increased and decreased size of temporal lobes in BPI patients, as well as no change (Beyer and Krishnan 2002). Findings are slightly more consistent when evaluating the amygdala and hippocampus. Two studies have found increased amygdala volumes Altshuler et al 2000, Strakowski et al 1999, but one study found decreased volumes (Pearlson et al 1997). The hippocampus was reported to be of equal size to those of control subjects (Beyer and Krishnan 2002).

Regional segmentation studies are rare in BPI. Zipursky et al (1997) studied white and gray matter volumes separately for frontal and frontal–temporal regions but found no differences compared with control subjects. Harvey et al (1994) found decreased gray matter volumes in temporal lobes and no difference in white matter volumes. They did not study the frontal lobe.

Two segmentation studies reported decreased overall gray matter volumes Lim et al 1999, Lopez-Larson et al 2002, whereas others found no difference compared with control subjects Brambilla et al 2001a, Sassi et al 2002, Schlaepfer et al 1994, Strakowski et al 1993, Zipursky et al 1997. No decrease in overall white matter volumes was found in five studies Brambilla et al 2001a, Lim et al 1999, Lopez-Larson et al 2002, Sassi et al 2002, Zipursky et al 1997, but Strakowski et al (1993) observed a decreasing trend in white matter volumes in first-episode mania patients. A consistent finding has been the association of BPI with increased occurrence of white matter hyperintensities. Bearden et al 2001, Benabarre et al 2002

Although MRI studies of BPI report somewhat mixed results, the emerging pattern seems to be consistent with a neuroanatomic model of mood regulation (Beyer and Krishnan 2002). The causal relationship between these findings and bipolar disorder is still to be resolved. It has been proposed that the findings represent neurodevelopmental abnormalities, originating either from genetic or environmental risk factors. In both cases, we would expect to see alterations already at a very early stage of the disorder. Hirayasu et al (1999) reported a decrease in subgenual cingulate cortex volumes in first-episode manic patients with a family history of BPI. Strakowski et al (1993) observed several structural brain abnormalities in first-episode mania. Friedman et al (1999) found frontal sulcal increase in adolescent BPI patients. Several studies have shown white matter hyperintensities already in children and adolescents with BPI Botteron et al 1995, Lyoo et al 2002, Pillai et al 2002.

A valid method for further examining the role of neurodevelopmental defects in bipolar disorder would be to investigate relatives of BPI patients. This would also offer the possibility of dissecting genetic factors predisposing to the disorder, a rational notion to emerge from studies showing high heritability for bipolar disorder Cardno et al 1999, Kendler et al 1995. Magnetic resonance imaging studies of relatives of bipolar patients are rare. An MRI study of six discordant monozygotic twin pairs with six control twin pairs was recently performed by Noga et al (2001). They measured basal ganglia, the amygdala–hippocampus, and cerebral hemispheres and found larger caudate nuclei and no typical asymmetry of hemispheres among healthy co-twins of bipolar twins compared with control subjects. One study (Ahearn et al 1998) examined white matter hyperintensities in a family of 21 members with a high loading of BPI disorder and found a high prevalence of findings in both affected and unaffected family members.

In the present study, we performed MRI scans of the brain on a nationwide sample of twins with bipolar disorder, their co-twins, and a demographically balanced sample of control twin subjects, with careful consideration of confounding factors. Based on previous findings, we decided to explore frontal and temporal gray and white matter volumes and regional sulcal and ventricular volumes. Fluid volumes were included because they can offer information about adjacent loss in brain tissue (Symonds et al 1999). The established high heritability of bipolar disorder Cardno et al 1999, Kendler et al 1995 makes it reasonable to postulate that subjects at high genetic risk also show detectable biological signs of increased liability to the disorder. Here, we hypothesized that any brain structural alterations detected in bipolar twins would also occur in healthy co-twins but to a lesser extent.