Risk Prediction Model And Neurobiological Markers Of Clinical High Risk For Psychosis

[Background & Objectives] The probability of clinical high risk for psychosis transition into psychotic disorders is about 30%.Early intervention can reduce the risk of CHR transition.However,at present the intervention is for the whole of the CHR population,rather than CHR with the highest risk of transition,because it is still based on clinical assessment to identify,while simply clinical assessment is difficult to predict CHR transition or not.Therefore,the aims of this study include: 1.to construct the risk predictive model for CHR transition combining clinical assessment and cognitive function test;2.to explore the underlying pathological mechanism of CHR transition by comparing the hippocampal subfields among CHR with different clinical outcomes;3.to explore the feasibility of nutritional intervention by comparing erythrocyte membrane unsaturated fatty acids among CHR with different clinical outcomes.[Methods] At the baseline,the CHR was screened using the Prodromal Questionnaire,Breif version(PQ-B)/ PRIME Screening Scale and the Structured interview for psychosis high risk syndrome(SIPS).Meanwhile,the healthy control subjects(HC)were recruited with matching demographic data to CHR.Once the subjects were enrolled,the cognitive function was evaluated using the MATRICS Consensus Cognitive Battery(MCCB),blood specimen was collected and brain structure images were scanned.Regular follow-up was conducted for all subjects mainly by face-to-face interviews,followed by telephone,short messages,or access to outpatient medical records,to assess the clinical outcome.During follow-up,the clinical outcome of CHR was assessed with SIPS,and CHR were divided into conversion,non-remession,and remession,and HC still belonged to healthy population.According to the purpose,the study was divided into three parts.The first part was used to construct the multivariate risk predictive model for CHR transition combining clinical symptoms and neurocognitive function by Cox regression analysis.The second part introduced the factor analysis to reduce the variance of the variables(for 24 hippocampal subfructures),and compared the hippocampal subfields volumes among CHR with different outcomes by multivariate analysis of variance,and analyzed the predictive effect of the different regions of hippocampus on CHR transition.The third part compared the content of polyunsaturated fatty acids in erythrocyte membrane among CHR with different clinical outcomew,and explored the correlation between long chain polyunsaturated fatty acids and CHR transition.[Results] The first part of the study was conducted with 224 CHR and 127 HC.Through an average of 2.3 years` follow-up,the main clinical outcome of CHR was transition to psychosis,non-remission and remission;there were 14(4.91%)CHR lost.First,we selected 15 potential risk factor including Father`s education,Mother`s education?family`s income,P2,P5,N1,N2,N3,N5,D4,GAFd,TMT,and BVMT;Secondly,151(71.90%)CHR were randomly selected by the complex sampling with stratification according to year of recruitment to construct the Cox regression risk model,while the other 59 samples were used to validate the predictive model.ROC curve analysis showed the variables of P2,N5,TMT,and BVMT to be moderately predictive,with AUC 0.724,and the sensitivity and specificity were 67.74% and 72.50%,respectively.The model is validated having good predictive performance,with AUC of 0.737.The second part introduced the factor analysis to reduce the 24 hippocampal subfields to seven common factors.And then the main effects of clinical group on hippocampal subfields in units of the seven common factors were observed by multivariate analysis of variance(ANOVA).Thus the left hippocampal fissure,right CA1,right hippocampal fissure,and righ parasubiculum were selected as regions of interest(ROI),and the right hippocampal fissure finally entered into the Cox regression risk model with AUC 0.816,which is comparable to the previous model with AUC 0.752(p <0.05),improving the predictive efficacy of the model significantly.In the third part,there were 100 samples,including 70 CHR and 30 HC.We compared the content of polyunsaturated fatty acids in erythrocyte membrane among CHR with transition,non-remission and remission by multivariate analysis of variance.C22: 4n6 in CHR group was significantly lower than that in HC group,while C20: 5n3 was significantly higher in the HC group.In addition,the ratio of n6 to n3 in the group who have transitioned to psychosis was significantly larger than that in the remission group.The ROC curve were used to set the cutoff of the ratio of n6 to n3 polyunsaturated fatty acids in the erythrocyte membrane.The results showed that the number of CHR who have transitioned is larger in the group with high ratio than that with low ratio.[Conclusion] 1.The validity of the predictive model for CHR transition to psychosis can be improved by combining clinical symptoms with neurocognitive function.2.There are abnormalities in the hippocampal subfields before the occurrence of psychosis,which may be due to neurodevelopmental abnormalities and imflammation;and right hippocampal fissure can improve the predictive validity of the risk model for CHR transition.3.n6 and n3 polyunsaturated fatty acids may be potential risk factor for CHR transition;n3 PUFAs can regulate the ratio of n6 to n3 on the one hand,and have anti-inflammatory effect on the other hand,and thus supplement of n3 PUFAs may be one of nutritional intervention means to reduce the CHR conversion rate.