1. Analysis Of Pathogen Distribution And Drug Resistance Of Nosocomial Infection In Patients With Severe Trauma2.Predictive Value Of STREM-1in The Treatment Response To Severe Trauma With Nosocomial Infection

Background Trauma was the main cause of death in Shenzhen especially outsidethe second frontier since2001. The incidence of nosocomial infection in patients withsevere trauma was high at61.1%. Severe infection was inclined to develop into sepsisand multiple organ dysfunction syndrome (MODS), which would be the main causefor death in later stage. As far as we know, administrating antibiotics to the pateintswith severe infection immediately and properly would help to improve survival. Butantimicrobial therapy poses an important clinical challenge by the diversebacteriology of complicated infections and the emergence of bacterial resistance. Sothe information about the spetrum of bacteria and drug sensitivity is very importantfor us to start empiric antimicrobial therapy.Objective To provide data for empirical antibiotic therapy through analyzing thepathogen distribution and drug resistance of nosocomial infection in patients withsevere trauma.Methods1. Specimens, such as sputum, urine, blood, secretion of wound surface,and drainageof Body cavity etc, of131patients with severe trauma admitted into intensive careunit from January2010to October2011were collected.2. Strains of bacterium in the specimens were identified and antibioticsensitivities were detected by VITEK-32automatic bacterium analyses system.3. Sex, age, the main site of injury, injury severity scale(ISS), acute physiology andchronic health evaluation(APACHE II)and the occurrence time of infection wererecorded to analyze the infection rate, the infection site，the pathogen distribution anddrug resistance and the risk factors of nosocomial infection.Results1.139specimens were identified containing pathogens and the incidence ofnosocomial infection was51.87%. In total191strains， the gram-negative bacteria accounted for61.26％(117/191,), dominated mainly by pathogens includingKlebsiella（30/191,15.71%）, Escherichia Coli（24/191,2.57%）, Acinetobacterbaumanni I （20/191,10.47%）, Pseudomonas aeruginosa （18/191,9.42%）,Enterobacter（15/191,7.85%） and Stenotrophomonas maltophilia（5/191,2.62%）.The gram-positive bacteria accounted for38.74％(74/191), mainly includingStaphylococcus aureus（30/191,15.71%）, coagulase negative staphylococcus (CNS)（20/191,10.47%）and Enterococci（16/191,8.38%）. No fungi was isolated. Mixedinfection rate was35.3％.2. The major pathogens identified from specimens of sputum, secretion of woundsurface,urine and blood were Klebsiella, Staphylococcus aureus, Escherichia Coliand CNS, respectively.3. Drug resistance to common antibioties: The positive rate of ESBLs in Klebsiellaand Escherichia Coli was61.23%and65.82%, respectively. The resistance rate ofKlebsiella and Escherichia Coli for imipenem was3.3%～4.2%, forpiperacillin/tazobactam and cefoperazone/sulbactam was6.7%～16.7%, for amikacinwas8.3%～13.3%and for Third generation cephalosporins was36.7%～70.8%. Theresistance rate of Acinetobacter baumanni I for imipenem was10%, forcefoperazone/sulbactam was25%and for other antibiotics was above45%. Theresistance rate of Pseudomonas aeruginosa for imipenem was16.7%, for amikacinwas16.7%, for piperacillin/tazobactam was27.8%, and for other antibiotics wasabove30%. No strains of Staphylococcus aureus, CNS and Enterococci were found toresist to vancomycin and Linezolid. But other antibiotics were found to have above50%rate in resistant to vancomycin and Linezolid.4. The rate of nosocomial infection in the pateints with severe trauma was61.07％.The respiratory tract was the most common site for infection. The occurrence time ofinfection was5.48±3.22days after injury.5. ISS and APACHE II in the infection group were significantly higher than those inthe non-infection group (26.66±9.44vs.20.68±4.43and19.99±8.86vs.16.08±7.94, P<0.05, respectively). The constituent ratio of brain injury and thoracic injury ininfection group were significantly higher than those in non-infection group (P <0.05, respectively). The constituent ratio of the limbs injury in non-infection group wassignificantly higher than those in infection group (P <0.05).Conclusion1. The incidence of nosocomial infection in patients with severe trauma was high andrelated to the severity of injury. The respiratory tract was the most common site forinfection.2. The conmon strains of gram-negative bacteria in pateints with infection after severetrauma included Klebsiella, Escherichia Coli, Acinetobacter baumanni I andPseudomonas aeruginosa. The positive rates of ESBLs for Klebsiella and EscherichiaColi were high. The resistance rates of Klebsiella and Escherichia for carbapenem andCompound Preparations of β-Lactamase Inhibitors were low. The resistance ratesof Acinetobacter baumanni I and Pseudomonas aeruginosa for carbapenem were low.3. The resistance rate of coccus for oxacillin was high. But there is no strains ofcoccus was found to resist to vancomycin and Linezolid.. Background Empiric antimicrobial therapy is important in the beginning of treatmentto severe trauma with nosocomial infection. It is also important to turn to targetedtherapy within48-72h according to germiculture and treatment response. However,there are no reliable markers of nosocomial infection to judge the response to initialtreatment. SIRS criteria such as fever, tachycardia, and tachypnoea, are observed inmost intensive care patients and are fairly nonspecific features of various underlyingdiseases. C-reactive protein (CRP) is considered to be an early nonspecific butsensitive marker of inflammation. Evidence has shown that CRP was not only goodfor diagnosis, but also as an indicator of infection in severe trauma patients.Procalcitonin (PCT) is a rather specific marker for severe bacterial infection has beenused to guide the initiation of antibiotic treatment in patients presenting withsuspected sepsis. However, Tang et al. demonstrated that PCT cannot reliablydifferentiate sepsis from SIRS in critically ill adult patients. Plasma concentration ofsoluble triggering receptor expressed on myeloid cells-1(sTREM-1) has been shownto be helpful in identifying critically ill patients suffering from infection. Gibot et al.found that the levels of sTREM-1in plasma was more helpful than CRP and PCT indifferentiating patients with sepsis from those with SIRS. It was reported thatsTREM-1higher than40pg/mL had sensitivity56.5%and specificity91.7%for thedifferential diagnosis between SIRS and sepsis after multiple injuries. If sTREM-1ismore effective in predicting treatment response to severe trauma with nosocomialinfection than CRP or PCT is still need further investigation.Objective To investigate the predictive value of sTREM-1in the treatment response tosevere trauma with nosocomial infection, we observed the plasma levels of sTREM-1,PCT and CRP before and during empiric antimicrobial therapy and analyzed therelationship between the plasma levels of sTREM-1, PCT and CRP and the outcomesof treatment.Methods 1. Eighty severe trauma patients with nosocomial infection admitted into intensivecare unit from January2010to October2011were enrolled in the study and receivedempirical antibiotic therapy according to antimicrobial treatment guidelines,knowledge of bacterial resistance in the community and the severity of the illness.2. Levels of serum sTREM-1,PCT and CRP were measured by sandwichenzyme-linked immunosorbent assay, immunofluorescence and turbidimetric assayrespectively on day0,day1,day2and day3.3. Sex, age, the main site of injury, injury severity scale(ISS), acute physiology andchronic health evaluation(APACHE II), mechanical ventilation time≥48h, inadequateantibiotic treatment, ICU stay (d) and mortality were recorded. The objects weredivided into two groups according to treatment response.4. The time course of sTREM-1,PCT and CRP plasma levels was proceeded repeatedmeasurement data F-test. The plasma levels of sTREM-1, PCT and CRP at day3topredict nonresponse was managed by receiver operating characteristic(ROC). The riskfactors were analyzed by univariate and multivariate logistic regression analysis.Results1.The serum levels of sTREM-1, PCT and CRP of all patients changed significantlybetween pre and post antibiotic therapy（P＜0.001）. In the response group, sTREM-1,PCT and CRP decreased significantly at different time points（P＜0.001）. In thenonresponsive patients, PCT and CRP also decreased, while sTREM-1remainedstable （P>0.05）. There are no differences between the response and nonresponsegroups in the serum levels of sTREM-1,PCT and CRP on day0. The levels of PCT inthe response group was significantly lower than that in the nonresponse group at day1,day2and day3. The levels of sTREM-1and CRP in the response group wassignificantly lower than that in the nonresponse group at day2and day3.2. Area under ROC curve of sTREM-1, PCT and CRP to predict nonresponse was0.880、0.865and0.797respectively.3. Univariate logistic regression analysis indicated that brain injury, ISS, APACHE II,mechanical ventilation time≥48h, the levels of sTREM-1, PCT and CRP at day3 were involved in predicting nonresponse（P<0.05）.4. Using multivariate logistic regression analysis, APACHE II and day3PCTpredicted treatment response significantly and independently（P<0.05）.Conclusion1. The failure rate of initial empiric antibiotic treatment is high in severe traumapatients with nosocomial infection. Failure in the initial empiric antibiotic treatmentwill significantly increase the length of ICU stay and mortality.2. Brain injury, ISS, APACHE II, mechanical ventilation time≥48h, the serurm levelsof sTREM-1, PCT and CRP on day3could be used to predict treatment response inpart. Both APACHE II and PCT on day3could be useful in predicting treatmentresponse independently.3. The prognostic performance of sTREM-1was not superior to that of PCT. sTREMwas found to be an adjuvant biomarker to PCT to detect nonresponsive patients early.