Role Of Neutrophil To Lymphocyte Ratio、Platelet To Lymphocyte Ratio And C-reactive Protein To Albumin Ratio In Sepsis Severity Evaluation And Prognosis

Objective:To explore the association among neutrophil to lymphocyte ratio(NLR),platelet to lymphocyte ratio(PLR),C-reactive protein to albumin ratio(CAR)and30-day mortality,as well as diagnosis and severity assessment of sepsis.Moreover,to establish combined prediction models in order to evaluate and verify their predictive efficacy on 30 day mortality of septic patients.Method:This is a single center,retrospective study.From July 2020 to January 2021,549 patients with initial diagnosis of "infection" in the First Affiliated Hospital of Fujian Medical University were enrolled.241 patients were excluded.The remaining 308 patients were divided into groups according to the severity of illness / sepsis and 30 day survival.NLR,PLR and CAR were divided into quartiles in the former group,and those with significant differences in univariate analysis were included in logistic regression analysis,In the former,NLR,PLR and CAR were grouped according to quartiles,and covariants with significant differences in univariate analysis were included in logistic regression analysis.The latter,covariants with significant differences in univariate analysis were included in logistic regression analysis;In addition,before building predictive models based on Logistic regression equation,and then drawning ROC curve for sepsis patients to further access the difference in predictive efficiency of 30-days survival rate between prediction models and qSOFA.Results:1.In this study,549 patients considered as "infection" were included,241 patients were excluded.The rest patients were divided into three groups according to the severity of infection: 74 patients with general infection(30 day mortality = 4.05%),97 patients with sepsis(30 day mortality = 8.25%)and 137 patients with septic shock(30 day mortality = 45.99%),There were significant differences among the three groups in vital signs and consciousness,recent operation history,inpatient department,infection site,etiology,advanced support treatment and 30-day survival rate(P < 0.050),SOFA,APACHE II,lactate level,PCT increased,albumin decreased and high level of CAR were assosiated with sepsis(P < 0.100).However,only SOFA and blood lactate level were the independent risk factors for sepsis after adjusting for confounding factors(P < 0.05).2.Sepsis and septic shock Patients were compared by univariate analysis,there were significant differences in age,heart rate,respiratory rate,MAP,state of consciousness,recent operation history,inpatient department,infection site,etiology,qSOFA,SOFA,APACHE II score,blood lactate level and advanced life supports during treatment between the two groups(P < 0.050).By logistic regression analysis,it was found that higher SOFA score,blood lactate level and heart rate(all P < 0.050)were independent risk factors for septic shock.3.Patients were divided into survival group(234 cases)and death group(74 cases)according to surviving or not after 30 days.There were significant differences in age,recent operation history,mean arterial pressure,respiratory rate,body temperature,state of consciousness,inpatient department,length of hospital stay,advanced support treatment,albumin level,SOFA,qSOFA and APACHE II score between the two groups(P < 0.050),In addition,the increased CAR level was associated with poor prognosis(P < 0.100).4.Logistic regression equation was used to establish the combined prediction model of sepsis prognosis,and the ROC curve was drawn: qSOFA + Lac + NLR(AUC: 0.697,95% CI: 0.642-0.747),qSOFA + Lac + car(AUC: 0.701,95% CI:0.647-0.752),qSOFA + Lac + NLR + CAR(AUC: 0.705;AUC: 0.697,95% CI:0.642-0.747);95% CI 0.641-0.756),and the predictive efficacy of the three models was better than that of qSOFA(all P < 0.001).5.By drawing the survival curve,further explored the predictive ability of NLR and CAR for the prognosis of sepsis according to their quartile stratification.The results showed that NLR and CAR had no significant assosiation with the poor prognosis of sepsis.Cox proportional regression model showed that SOFA,APACHE II score and increased blood lactate level were independent risk factors for poor prognosis of sepsis patients(P < 0.050).Conclusion:1.NLR and CAR are good predictors of poor short-term prognosis in patients with infection.2.NLR,PLR and car have no significant value in accessment of the severity and prognosis of sepsis.High SOFA and APACHE II score were independent risk factors for poor prognosis in patients with sepsis.3.The combined prediction model S: qSOFA +Lac + NLR,qSOFA + Lac + CAR,qSOFA + Lac + NLR + CAR were better than qSOFA in predicting the prognosis of sepsis.The high NPV of qSOFA + Lac + NLR was more reliable than SOFA to support the use for a safe ruling out of infected patients at high risk of mortality at admission.