Biomarkers In Predicting Development And Outcome Of Acute Respiratory Distress Syndrome

Background and objectives:ARDS is a type of acute diffuse,inflammatory lung injury^[1].The incidence rate of ARDS admitted to ICU is 10.4%^[2].Although we use lung protective ventilation,the mortality of ARDS in hospital and intensive care unit is still more than 40%^[3].At present,there are no specific drugs that are confirmed to be effective except lung protective ventilation with low tidal volume and restrict fluid administration.To prevent development and progression of ARDS are key factors to reduce mortality of ARDS.Thus,its of great value to study factors associated with development of ARDS and poor outcome of ARDS.That would help us to recognize high risk factors of ARDS and to take effective intervention to prevent development of ARDS.MODS is often accompany withARDS.Both have the same mechanism,which is imbalance of proinflammatory response and anti-inflammatory response.The morphological hallmark of the acute phase of ARDS is diffuse alveolar damage^[1,4].The plasma levels of SP-D increase early in ARDS may reflect pulmonary epithelial injury and consequent increased permeability^[5].Interleukine-8 is the major chemokine of neutrophil.Decoy receptor 3 is also knowed as Tumor Necrosis Factor,Member 6B,which can inhibit cell apoptosis and has a function of immunomodulation^[6].Pneumonia is the most common risk factor for the development of pulmonary ARDS^[7].The limitation of the current definitions of acute respiratory distress syndrome and the heterogeneous character may be reasons for failure of therapeutic interventions^[7,8].So we focus on predicting development of pulmonary ARDS.We evaluate the value of serum SP-D,IL-8,DcR3 in predicting development and outcome of acute respiratory distress syndrome and thus provide theory for biomarker guided treatment.We intended to establish a predictive model for development of pulmonary ARDS.Methods:Part one:Biomarkers in predicting development of pulmonary ARDSWe conducted a prospective nested case-control study to evaluate serum SP-D,IL-8and DcR3 in predicting development of pulmonary ARDS and to establish a predictive model for development of pulmonary ARDS.96 at-risk patients for pulmonary ARDS were included.Patients who had pneumonia with respiratory failure were analysized as a subgroup.The primary outcome was development of pulmonary ARDS in a week.Patients who developed ARDS were divided to developed group,while patients who did not developed ARDS were divided to non-developed group.Pulmonary ARDS was diagnosed according to the Berlin Definition^[1].Sera SP-D,IL-8 and DcR3 were measured.Baseline sera SP-D,IL-8 and DcR3 were compared between groups.ROC analysis were used to evaluate predictive value.Multivariate Logistic-regression were used to establish a prediction model for development of pulmonary ARDS.Part two:Biomarkers in predicting outcome of ARDSStudy design was similar to part one.There were two sections.In the first section,13heterogeneous pulmonary ARDS were studied.The primary outcome was 30 day all-cause mortality.The secondaryoutcome was whether PaO₂/FiO₂decreased.Patients were divided to survival group and non-survial group or progress group and non-progress group.Serial sera SP-D,IL-8 and DcR3 levels were studied.In the second section,42 ARDS patients were studied.Multivariate Cox-regression was used to evaluate the assotiatioinof serum DcR3with30 day all-cause mortality of ARDS.Results:Part one:1.6.25%patients developed pulmonary ARDS.There was no difference between sex.Patients were younger in developed group?P=0.040?,while baseline APACHE?scores were higher in developed group?P=0.039?.Subgroup analysis showed that 15%patients developed pulmonary ARDS in pneumonia with respiratory failure.APACHE?scores were of no significant statistic difference bwtween the two group.2.Baseline sera SP-D were higher in developed group?P<0.05?,while baseline sera IL-8 and DcR3 were of no significant statistic difference?P>0.05?.Subgroup analysis showed that baseline seraSP-D werestill higher indeveloped group?P<0.05?.3.ROC analysis showed that SP-D?AUC=0.796?was superior than APACHE?score?AUC=0.721?in predicting development of pulmonary ARDS.Besides,sensitivity of SP-D and APACHE?score were 83.3%and 69.7%.4.Logistic regression showed that after adjustment for age andAPACHE?score,risk of developing pulmonary ARDS increased with higher baseline SP-D?adjusted OR=1.118,95%CI:1.028-1.216?.Subgroup analysis showed that SP-D were still an independent predictor of ARDSp?adjusted OR=1.182,95%CI:1.004-1.390?.5.ARDSp prediction model-1:logit?P?=-0.019+0.244×APACHE?score+0.111×baseline SP-D-0.145×age.Model-2:logit?P?=2.296+0.379×APACHE?score+0.167×baseline SP-D-0.230×age.These first model showed good performance in predicting ARDSp with sensitivity of 83.3%,with specificity of 93.3%,with AUC of 0.955?95%CI0.908-1?,respectively.Part two:1.In the first section,30 day all-cause mortality was 53.8%.There was 80.0%death in the progress group.In the second section,30 day all-cause mortality was 50.0%.40.5%patients developed MODS.2.Baseline sera SP-D,seraDcR3 and sera interleukin 8 in progress groupwere not significantly different from non progress group.Sera SP-D,DcR3 and IL-8 levels increased over time in the progress group?P value were 0.034,0.047 and 0.238?,but remained in a low level in non-progress group?P>0.05?.3.When measured after baseline,sera SP-D,DcR3 and IL-8 were higher in nonsurvivors than survivors?P value were 0.029,0.040 and 0.423?.4.Spearman rank correlation coefficent of DcR3 to IL-8 and APACHE?score were0.718 and 0.472.The correlation coefficent between DcR3 and IL-8 increase to 0.809 in nonsurvavors.5.Patients who developed MODS had higher baseline DcR3,P=0.011.6.Multivariate Cox regression showed that baseline sera DcR3 more than2.7ng/m L?HR=2.557,95%CI:1.016⁶.432,P=0.046?,developed MODS?HR=4.667,95%CI:1.751¹2.443,P=0.020?and PaO₂/FiO₂ less than 165mmHg?HR=2.737,95%CI:1.049⁷.141,P=0.040?were independently associated with 30 day all-cause mortality for ARDS and contribute to prediction model.Conclusions:1.Sera SP-D may be a useful biomarker to predict development of pulmonary ARDS,while values of sera DcR3 and IL-8 are limited.Higher baseline sera SP-D and increase overtime indicate higher risk of developing pulmonary ARDS.2.Changes of sera SP-D,IL-8 and DcR3 are associated with progression of pulmonary ARDS.These three biomarkers increase overtime indicate poor outcome of ARDS and may contribute to biomarker guided treatment.3.DcR3 is an independent predictor of 30 day all-cause mortality in ARDS.Besides,DcR3is associated with imbalance of proinflammatory response and anti-inflammatory response as well as development of MODS.