Prediction of functional outcome 18 months after a first psychotic episode: a proton magnetic resonance spectroscopy study
Wood SJ, Berger GE, Lambert M, Conus P, Velakoulis D, Stuart GW, et al.;
Commented by , 30 Sep 2006
Aim of the study
It is a well established fact that people with schizophrenia have enlarged brain ventricles compared to normal controls. It is also clear that the brain structure of people with schizophrenia is already different to that of normal controls already in the first episode.
Apart from these rather global findings, the results of neuroimaging research attempting to measure brain volumes to predict clinical outcome years later in prodromal stages and early stages of schizophrenia have been inconsistent. Proton magnetic resonance spectroscopy is a relatively new method which has not been used much in schizophrenia.
In contrast to conventional imaging it measures the concentration of various brain metabolites and may therefore better reflect the schizophrenic pathology than simple volumetric measures. The authors, therefore, investigated whether spectroscopic measurement of frontal and temporal lobes can predict functional mid-term outcome in patients with first-episode psychoses.
Method
The authors examined 46 patients from a specialised early psychosis unit with a first-episode of psychosis using proton magnetic resonance spectroscopy. They calculated the ratio of N-acetylaspartate (NAA) and choline containing compounds to creatine and phosphocreatine (Cr) (NAA/Cr ratio) by examining left prefrontal cortex and left mediotemporal lobe voxels.
NAA is considered to be a marker of cellular integrity which is thought to be reduced in conditions with brain cell loss. The prefontral region was chosen, because previous studies in schizophrenia had shown that NAA is most reduced here. To predict functional outcome at 18 months after admission they entered these parameters together with age and duration of untreated psychosis in a regression analysis.
Functional outcome was defined as scores on the Global Assessment of Functioning Scale (GAF), Clinical Global Impression scales (CGI), Social and Occupational Functional Assessment Scale (SOFAS) and the number of admissions during the analysis period. They analysed the impact of premorbid function and negative symptoms (measured by a subscore of the Positive and Negative Syndrome Scale, PANSS) in secondary analyses.
Results
The only imaging result that predicted outcome was the degree of NAA/Cr ratio in the prefrontal cortex. Poorer outcome on all measures (GAF, CGI, SOFAS, number of admissions) was predicted by a low NAA/Cr ratio. 17% to 30% of the outcome variance was explained by the frontal NAA/Cr ratio. Negative symptoms and premorbid functioning improved the predictive power of the model while other PANSS factors did not.
Dr Leucht's comments
The authors’ findings are extremely interesting in the context of imaging research in early psychosis. A general difficulty of this area is that schizophrenia and psychosis in general are very heterogeneous disorders. Imaging in this area is still at an exploratory stage, in part because the pathomechanism behind psychosis is not clear yet.
I feel that a finding explaining 17% to 30% of the outcome variance is therefore quite remarkable, although replications are of course needed. The authors discuss that conventional volumetric measurements in this area were usually negative, possibly because in contrast to their spectroscopic method the measurement of simple volumes may not reflect the underlying pathologies.
The authors’ main conclusion is that prefrontal neuronal dysfunction is a predictor of poor mid-term (here 18 months) prognosis rather than a trait marker of first episode of schizophrenia. They express the hope that this finding may be used to guide treatment, but that further research is needed. While a lot of work is still needed in this area, I believe that it is an important path of research.