Regional change in brain morphometry in schizophrenia associated with antipsychotic treatment

Abstract

The purpose of this pilot study was to: (1) determine if regional brain volume change occurs in schizophrenia patients during very short periods of withdrawal from, or stable treatment with, antipsychotics, and; (2) compare results of region-of-interest (ROI) to voxel-based morphometry (VBM) methods. In two small groups of schizophrenic inpatients, magnetic resonance imaging was performed before and after antipsychotic withdrawal, and at two time points during stable chronic antipsychotic treatment. Regional brain volumes were measured using ROI methods. Grey matter volume was measured with VBM. The medication withdrawal group showed no effect of treatment state or antipsychotic type on regional brain volumes with ROI analysis, but effects of both treatment state and antipsychotic type on grey matter volume were observed with VBM in right middle frontal, right medial frontal, right and left superior frontal, right cingulate, and right superior temporal gyrii as well as in the right and left hippocampal gyrii. The chronic stable treatment group showed an effect of time on right caudate, left hippocampal, and total cerebrospinal fluid volumes with ROI analysis, while effects of both time and antipsychotic type were observed with VBM on grey matter volume in the left superior temporal lobe. No findings survived correction for multiple comparisons. A positive correlation between regional volume change and emerging psychopathology was demonstrated using ROI methods in the medication withdrawal group. Treatment state and emergent symptoms in schizophrenia patients were associated with regional volume change over very short time periods. Longitudinal regional brain volume change in schizophrenia patients is likely physiologic and therefore potentially reversible.

Introduction

Investigators have detected basal ganglia volume change in schizophrenic patients with magnetic resonance imaging (MRI) after long antipsychotic treatment periods. Increases of 5.7% (Chakos et al., 1994) and 15% (Keshavan et al., 1994) in caudate volume were demonstrated in schizophrenic patients after 10 months and 4 years of typical antipsychotic treatment, respectively, in longitudinal MRI studies (DeLisi et al., 1995, Corson et al., 1999, Lieberman, 1999). Typical and atypical antipsychotics appear to affect basal ganglia volume differently (Chakos et al., 1998). Decreases in basal ganglia volume are generally associated with atypical antipsychotic treatment. Decreases in caudate volume have been associated with clozapine treatment over mean periods of 24 (Scheepers et al., 2001) and 54 weeks (Chakos et al., 1995). After a mean treatment period of 2 years with atypical antipsychotic medications—clozapine, risperidone and olanzapine—decreases in basal ganglia volume were observed (Corson et al., 1999), although one cross-sectional study suggests caudate volume can increase after atypical antipsychotic treatment (Vetsa et al., submitted for publication). A morphometric analysis of Vetsa et al.'s cross-sectional sample showed altered caudate shape in both first-episode and chronic schizophrenia patients, after months of atypical antipsychotic treatment. Treatment with typical antipsychotics is usually associated with increased basal ganglia volume. The exact physiological mechanism of basal ganglia volume change associated with antipsychotic treatment is unknown, but may be mediated by dopamine D2 receptor blockade (Corson et al., 1999) or other pharmacological properties (Chakos et al., 1994, Andersson et al., 2002) of these medications.

Progressive brain volume change in schizophrenic patients has been reported. Increases in the lateral ventricles and cerebrospinal fluid, and decreases in regional brain volumes, have been detected over 1–10 years, in adolescent patients with chronic illness (Rapoport et al., 1997, Jacobsen et al., 1998, Rapoport et al., 1999), in new-onset cases (DeLisi et al., 1992, DeLisi et al., 1995, DeLisi et al., 1997, Gur et al., 1998, Lieberman et al., 2001, McCarley et al., 2002, Gur et al., 2002, Cahn et al., 2002, Lieberman et al., 2005), and in chronic adult patients with schizophrenia (Nair et al., 1997, Davis et al., 1998, Mathalon et al., 2001). Furthermore, decreases in total frontal, temporal, and parietal cortical volume, exceeding those of normal controls, have been demonstrated in a longitudinal sample of childhood-onset schizophrenia when they entered adolescence (Sporn et al., 2003, Gotgay et al., 2004). Progressive volume change has been interpreted alternately as evidence of neurodegeneration or neurotoxicity in the schizophrenic brain (Lieberman, 1999, Mathalon et al., 2001, Jarskog et al., 2002, Jarskog et al., 2004, Christensen et al., 2004) or attributed to plastic, physiological adaptations of the brain to the environment, including pharmacologically-induced changes (Weinberger and McClure, 2002). The latter interpretation would suggest that these MRI volume changes are, at least potentially, reversible and not necessarily a reflection of structural change in neuronal populations.

Interest in the relationship between antipsychotic treatment for relatively long periods and volume change in regions besides basal ganglia followed the identification of progressive brain volume change in schizophrenia. Investigators examining the effects of antipsychotics for longer treatment periods on brain regions other than the basal ganglia report inconsistent results. Decreased whole brain grey matter volume has been observed after 3 months of treatment with haloperidol in first-episode schizophrenia patients (Lieberman et al., 2005). The opposite finding has also been reported (Nopoulos et al., 2005): total cerebral volume and grey matter volume increased over 2–5 years in patients treated with typical, rather than atypical antipsychotics as well as in normal controls. In a one-year follow-up study of first-episode schizophrenia patients (Cahn et al., 2002), no relationship was observed between cerebral grey matter change and antipsychotic type (typical, atypical, and clozapine). Decreased cerebral grey matter volume over time did correlate significantly with cumulative dose of antipsychotic medication. The relationship of antipsychotic treatment for longer periods on subcortical brain volumes is unexplored. No difference in hippocampal, caudate, prefrontal grey or white matter volumes were observed in chronic schizophrenia patients randomly assigned to haloperidol or clozapine for 10 weeks (Arango et al., 2003). In a cross-sectional study, treatment with atypical antipsychotic medications compared to haloperidol was associated with larger hippocampal volumes, even after controlling for differences in illness severity (Gerig et al., submitted for publication). However, this relationship held only for patients early in their illness, not for chronically ill patients. Differences in hippocampal shape were observed between schizophrenia patients treated with atypical and those treated with typical antipsychotics with the same data.

Few investigators have explored the relationship between regional brain volume change in psychiatric patients and brief medication treatment periods. Changes that occur over very short time periods are unlikely to reflect fixed structural abnormalities. Small increases (3%) in total grey matter volume have been detected after four weeks of treatment with lithium carbonate in patients with bipolar affective disorder (Moore et al., 2000). Larger (19%) decreases in thalamic volume have been observed in pediatric patients with obsessive–compulsive disorder after 12 weeks of treatment with paroxetine (Gilbert et al., 2000). A few investigators have examined brain volume change over very short antipsychotic treatment periods. Increases (4.2%) in cortical (frontal, parietal, temporal and occipital) grey matter volume were reported after 4 weeks of treatment with risperidone or ziprasidone in schizophrenia patients (Garver et al., 2005). In a similar study involving antipsychotic treatment over 4 weeks, no change in cerebral white matter volume was found (Christensen et al., 2004). During worsening of psychotic episodes, when patients were not taking antipsychotics, increases in total brain volume and decreases in lateral ventricular volume have been observed (Garver et al., 2000). The first objective of this study was to examine the effect of short-term antipsychotic treatment on regional brain volume in schizophrenia.

Methods employed for the analysis of longitudinal MRI data vary from one study to another. Voxel-based morphometry (VBM) has become popular because it minimizes rater biases implicit in region-of-interest (ROI) methods. The results VBM studies in schizophrenia vary considerably (Wright et al., 1999, Sowell et al., 2000, Wilke et al., 2001, Hulshoff et al., 2001, Thompson et al., 2001, Kubicki et al., 2002, Job et al., 2002, Ananth et al., 2002, Pantelis et al., 2002) and their interpretation can be difficult. Wright has proposed that VBM provides a reasonable estimate of regional brain volume and is useful for detecting small grey matter reductions located in areas with high variability of volume or inconsistent in location. It is unlikely that VBM will substitute for manual volumetry (Tisserand et al., 2002). VBM is vulnerable to partial volume effects when grey or white matter voxels are misclassified (Job et al., 2002). It alters location of maximal volume change by the use of smoothing (Kubicki et al., 2002). Validation of voxel-based studies has been limited to just a few brain regions, experimental conditions, and clinical populations. The secondary objective of this study was to compare ROI-based analysis to optimized VBM.

In this pilot study, regional brain volume changes were examined in schizophrenia patients undergoing a brief period of antipsychotic drug withdrawal. The subjects were a small group of schizophrenic patients who volunteered to undergo two structural MRI studies: the first scan was performed while patients were treated with typical or atypical antipsychotics; the second scan was performed after an average of twenty-five days on placebo. The comparison population consisted of schizophrenia patients who had remained on the same dose of the same antipsychotic medication for many months prior to the first scan. They underwent two MRIs after an average of fifty-two days. Standard ROI methods were compared to those obtained from optimized VBM.