Comparative Study To The Practice Of Scytokine And APACHE Scoring Systems In Predicting Severe Pneumonia Mortality In Elderly Patients

Objective: It’s crucial to take use of comprehensive assessment method in predicting severe pneumonia mortality and guiding the treatment in elderly patients,with complex pathogenesis,who have a high mortality rate seems.The study is divided into two parts:part one retrospective analysis of APACHE II score,APACHE IVscore,MEDS score and SIPF score in predicting severe pneumonia 28-day mortality in elderly patients,the aim of this study was to compare the effects of APACHE II score,APACHE IV score,MEDS score and shock index and SIPF score for the prediction of 28-day mortality in the elderly patients with severe pneumonia.Part two prospective analysis of APACHEII score, APACHEIV score and scytokine in estimating severe pneumonia mortality in elderly patients, the aim of this study was to compare the effects of APACHE II score,APACHE IV score and scytokine for estimating the prediction of 28-day mortality in the elderly patients with severe pneumonia.Method:Method of part one retrospective analysis:compared areas under the receiver operator characteristic(ROC) curves for 28-day mortality A total of 104 critical patients whose average age is 76.33±7.15 during June 2013 and June 2015,undertook APACHE II score,APACHE IV score, MEDS score and SIPF score.Correlations between the four scoring systems were compared according to the survival and death conditions in the 28-day observation period,then find out independent risk factor of mortality from Significant different variables by univariate and multivariate logistics regression analysis,the accuracy of evaluating the prognoses with four scoring systems were compared by areas under the ROC curve and Hanley-Mc Neil method.According to the cut-off calculate and compare the value of SMR. Method of part two prospective analysis: a prospective observational study was conducted at the central intensive care unit of Tianjin People’s Hospital during April 2015 and Novermber 2015,40 critical patients from whose average age is 76.70±6.95 were reviewed to identify the elderly patients with severe pneumonia undertook APACHE II score,APACHE IV score and scytokine on day-1,or day-3,or day-7,which were compared according to the survival and death conditions in the 14-day or 28-day observation.Correlations between two scoring systems and scytokine with the Pearson product-moment correlation coefficient,the comparison of TNF alpha, APACHE II score,APACHE IV score and 14-mortality or 28-mortality with the rank correlation coefficient, the accuracy of evaluating the prognoses with two scoring systems and TNF-α were compared by areas under the ROC curve and Hanley-Mc Neil method.Results: Results of part one retrospective analysis:Four scoring systems and age had good discrimination for predicting 28-day mortality in elderly patient with severe pneumonia,Univariate and multivariate logistics regression analysis find out APACHE II score and APACHEIVscore can predict mortality independently, with Regression coefficients(0.162,95%CI 1.037-1.334,P <0.05),(0.111,95%CI1.051-1.188,P < 0.05),(0.103,95%CI1.007-1.220,P <0.05);MEDS score,SIPF and age score don’t predict mortality independently, with Regression(0.484,95%CI0.225-1.688,P>0.05),(0.115,95%CI 0.716-1.109,P>0.05),(0.013,95%CI 0.097-1.220,P>0.05).The AUC of APACHE II score and APACHE IV score are 0.818(95%Cl 0.728-0.908),0.865(95%Cl 0.783-0.950).However,the APACHE II score and APACHE IV score did not have discrimination power for predicting mortality,with(Z=1.097,P>0.05).SMRof APACHE II score in the low-risk group is 0.95,while in the high-risk group is 1.33. SMR of APACHE IV score in the low-risk group is 1.06,while in the high-risk group is 2.14. Results of part two prospective analysis: APACHE II score,APACHE IV score and TNF-α were significant higher in non-suviors group on days 3,7(P<0.05),LI-1β and LI-10 levels have similar tendency in two groups(P>0.05).Pearman coefficient show that was significantly related, with γ(0.40,0.30,0.42,P<0.05);Spearman coefficient show that different variables were significantly positively related to 14-mortality,with γ(0.442,0.508,0.636,P<0.05);28-mortality γ(0.622,0.464,0.393,P<0.05).The AUC of TNF-α,APACHE II score,APACHE IV score,TNF-α combined with APACHEIV score and TNF-α combined with APACHE IVscore in the prediction of 14-mortality:0.802(95%CI0.609-0.904),0.792(95%CI0.651-0.933),0.731(95%CI0.564-0.897),0.806(95%CI0.669-0.943),0.815(95%CI0.682-0.947);The AUC of TNF-α,APACHE II score,APACHE IVscore,TNF-α combined with APACHEIV score and TNF-α combined with APACHEIVscore in the prediction of 28-day mortality:0.852(95%CI0.737-0.967),0.879(95%CI0.775-0.982),0.775(95%CI0.627-0.923),0.916(95%CI0.830-1.000),0.900(95%CI0.808-0.993).However,TNF-α combined with APACHE II score or APACHE IV score did not have discrimination power for predicting mortality(P>0.05).Conclusions:APACHE II score and APACHE IVscore can predict mortality in elderly patients with severe pneumonia independently.These data suggest that APACHE II,as well as APACHE IV score,is effective for predicting 28-day mortality in elderly patients with severe pneumonia,with the cut-off(26,60).but that the MEDS score and SIPF score does not have good discrimination and calibration for predicting mortality. These data suggest tha TNF-α,APACHEIIscore and APACHE IVscore are useful for predicting 28-day mortality,the higher TNF-α,APACHEIIscore and APACHE IVscore,the higher mortality APACHEIIscore and APACHE IVscore predict mortality in elderly patients with severe pneumonia.Compared to TNF-α, APACHE II score and APACHE IV score,TFN-α combined with APACHE II score or APACHE IV score have discrimination power for predicting 28-day mortality in elderly patients with severe pneumonia,Whether TNF-α combined APACHE II score or APACHE IV score makes no diffrences in predicting mortality.