Investigation Of Early Predictive Indicators And Methods For Post-traumatic Sepsis

BackgroundSepsis is one of the most common complications in the later stage of severe trauma,and it is the leading cause of in-hospital death for trauma patients.Early diagnosis of post-traumatic sepsis is essential for the clinical management.However,trauma patients are often in a state of“sterile inflammation”,which greatly affects the early diagnosis of sepsis.Due to the high incidence and mortality of sepsis after trauma,early recognition and individualized therapy for those patients at high risk are essential prerequisites to reduce the occurrence of sepsis in trauma patients.Therefore,identifying effective indicators and constructing an early prediction model of post-traumatic sepsis are highly warranted.The objective of this study is to identify indicators related to post-traumatic sepsis at three aspects:genetic background,clinical indicators and transcriptome level,and then construct prediction models of post-traumatic sepsis.Methods and materials1.Identification of genetic polymorphisms for sepsis susceptibility:On the one hand,according to the key words“Sepsis”and“Polymorphism”,literatures investigated the associations between genetic polymorphisms and sepsis susceptibility were identified by searching Pubmed,Embase,Web of knowledge and HuGe databases.Meta-analyses were carried out for the gene polymorphisms≥3 independent study groups,systematic evaluation was used to screen the gene polymorphisms<3 independent study groups.Those polymorphisms might have potential effects on the sepsis susceptibility were evaluated.On the other hand,on the basis of previous research,the clinical association studies,Western blotting,Luciferase reporter gene and Electrophoretic Mobility Shift Assay were applied to investigate the potential roles of PPARG rs10865710C/G in post-traumatic sepsis.2.Construction of the weighted genetic risk score（wGRS）model for post-traumatic sepsis:The above 64 polymorphisms related to sepsis susceptibility were genotyped in a Chinese trauma cohort with SNPscan methods.Based on the contribution of each polymorphism to post-traumatic sepsis,those genetic polymorphisms with an Average Decreasing Accuracy（MDA）≥1.0 by Random Forest were screened to construct the wGRS of post-traumatic sepsis.Meanwhile,the Area under the Curve（AUC）was applied to evaluate the prediction ability.Moreover,Net Reclassification Improvement（NRI）were used to evaluate the reclassification ability of wGRS for post-traumatic sepsis.3.Early prediction for the risk of post-traumatic sepsis:Clinical and laboratory variables of trauma patients within 24 hours after admission were collected,then the risk of post-traumatic sepsis was analyzed.Logical regression and Least Absolute Shrinkage and Selection Operator（LASSO）technique were adopted to identify variables contributing to the early prediction of post-traumatic sepsis.Then,the traumatic sepsis score（TSS）based on those selected variables was constructed.Moreover,the discrimination was assessed with the Area under the Curve（AUC）,and the calibration of the TSS were evaluated using the Hosmer-Lemeshow（H-L）goodness-of-fit test.4.Identification of transcriptional indicators associated with post-traumatic sepsis:On the one hand,lncRNA microarray was adopted to identify differentially expressed lncRNAs in peripheral blood mononuclear cells（PBMCs）induced by Lipopolysaccharides（LPS）.The GO and KEGG analyses were performed to investigate the functions of differentially expressed lncRNAs.Moreover,the lncRNAs-mRNAs co-expression network was constructed.We proposed to identify lncRNAs contributing to the incidence of sepsis.On the other hand,the public transcription database of trauma patients was mined to identify genes significantly differentially expressed after injury.Furthermore,the differentially expressed genes were verified in trauma patients.Additionally,the prediction abilities of those genes were investigated in severe trauma cohorts.Results1.Through systematic literature search,349 eligible articles on genetic associations of sepsis susceptibility were selected,involving 405 gene polymorphisms.Then,204 primary meta-analyses and 185 subgroup meta-analyses（age and ethnicity）were performed for 76polymorphisms with at least 3 independent studies.The results suggested 29 polymorphisms were associated with the incidence of sepsis.Meanwhile,329 polymorphisms with less than 3studies were systematically evaluated,and 63 polymorphisms were potentially associated with the risk of sepsis.Moreover,the clinical association studies verified the nuclear receptor PPARG rs10865710C/G was significantly associated with the risk of sepsis in trauma patients from Chongqing and Guizhou.Additionally,Western blotting,Luciferase reporter gene and Electrophoretic Mobility Shift Assay further indicated rs10865710C/G was a risk variation for post-traumatic sepsis by affecting the activity of PPARG enhancer and the ability binding to CREB2.2.A total of 64 polymorphisms associated with sepsis susceptibility were successfully genotyped in 883 trauma patients with SNPscan methods.Random Forest analyses indicated17 polymorphisms were with MDA≥1.0.Based on those 17 loci,the w GRS of post-traumatic sepsis susceptibility was constructed.The results demonstrated wGRS was significantly correlated with the risk of post-traumatic sepsis(OR=2.19（1.53-3.15）,P=2.01×10^-5),and the AUC of wGRS for the early prediction of post-traumatic sepsis was0.619（0.586-0.651）.When the wGRS was added to the clinical prediction model,the AUC of prediction model reached 0.768（0.739-0.796）,with an increase of 3.40%(P=8.00×10^-4).Moreover,the NRI of GRS was 25.18%（17.84%-32.51%）(P=6.00×10^-5).3.Totally,684 trauma patients,including 411（60.0%）patients in the training group and273（40.0%）patients in the validation group,were recruited.Based on Logical regression and LASSO analyses,7 variables（Injury severity score（ISS）,Glasgow Coma Scale（GCS）,Temperature（T）,Heart rate（HR）,Albumin（ALB）,International normalized ratio（INR）,and C-reactive protein（CRP））were selected from 50 clinical features to construct the TSS in training group.The results indicated the incidence of sepsis after trauma increased with an increasing TSS(P_trend=7.44×10^-2121 and P_trend=1.16×10^-13)in the training and validation groups.Meanwhile,the AUCs of TSS for post-traumatic sepsis were 0.799（0.757-0.837）and0.790（0.736-0.836）,which were superior to that of SOFA score（P<0.001）.Moreover,H-L test indicated TSS had good calibration ability（P=0.386 and P=0.082）.4.Analyses of lncRNA transcriptome data suggested 890 lncRNAs and 1635 mRNAs（FC≥2,FDR≤0.05）were differently expressed in PBMCs induced by LPS.Then,the GO and KEGG results indicated those lncRNAs were mainly associated with NF-KB,NLR,TLR and TNF signal pathways.These results might provide some potential lncRNAs for the diagnosis and treatment of sepsis.Meanwhile,the public database suggested VNN1 significantly increased after trauma,and VNN1 might be involved in the development and progress of traumatic sepsis.In two trauma groups including 283 and 121 patients,plasma vanin-1 were associated with post-traumatic sepsis(OR=3.92（2.68-5.72）,P=1.67?10^-12 and OR=4.26（2.22-8.17）,P=1.28?10^-5),and the prediction power of plasma vanin-1（0.82（0.77-0.87）and0.83（0.75-0.89））were significantly superior to that of CRP,PCT and APACHE II（P<0.05）.ConclusionIn the current study,a series of important polymorphisms related to sepsis susceptibility were identified through systematic review and meta-analyses.Then,the weighted genetic risk score of post-traumatic sepsis was constructed.Meanwhile,based on the clinical variables of trauma patients after admission,seven clinical features were selected to construct the traumatic sepsis score.Finally,some differentially expressed lncRNAs were identified to be new targets for sepsis research,and plasma vanin-1 may be a novel biomarker to predict the risk of post-traumatic sepsis.