| Sepsis is a systemic, deleterious host response to infection leading to severesepsis (acute organ dysfunction secondary to documented or suspected infection)and septic shock (severe sepsis plus hypotension not reversed with fluidresuscitation). Severe sepsis and septic shock are major cause of mortality inintensive care unit, the rapid diagnosis and management of which can influenceoutcome. Manifestations of sepsis could be serious, although they did notalways represent specific indicators of infection. In addition, noninfectiousclinical syndromes could exhibit sepsis-like symptoms, which not only making adiagnosis of sepsis more difficult, but also disturbing its therapy. Procalcitonin(PCT) and C-reactive protein (CRP) were main biomarkers to diagnose sepsis.PCT is released into the circulation approximately3–4h after exposure to anendotoxin, and reaches peak levels8–24h later. As a result, PCT has beenproposed to be a marker of bacterial infection in critically ill patients. CRP is aprotein that is synthesized by the liver in response to a stimulus that mainlyincludes IL-6. In the clinic, CRP is one of the most frequently assayedinflammatory markers in evaluating a patient’s response to antimicrobial therapy,diagnosing infection and for screening critically ill patients. To improve theaccuracy of diagnosing sepsis, monitoring of both novel biomarkers andtraditional markers of sepsis had been proposed. For PCT and CRP, limited datawere available regarding the use of both biomarkers for an evaluation of sepsisamong critically ill adult patients and less information about C-reactive proteinfor guiding antibiotic therapy in critically ill patients were found. Therefore, the objective of this study was to evaluate the usefulness of assaying PCT, CRP andthe combination of them for the identification of sepsis, and explore the value ofC-reactive protein for guiding antibiotic therapy in critically ill adult patients.Ⅰ C-reactive protein(CRP) serum levels in terms of their valuefor sepsis diagnosis and prognosisA retrospective cohort study was performed at the First Hospital of JilinUniversity (Changchun, China) using data obtained from critically ill patientstreated between January1and December31,2012. Patients18years or olderwith initial CRP levels recorded in their medical histories were included in thisstudy. Patients with unclear diagnoses, and patients who were hospitalized lessthan72h, were excluded.109critically ill patients were included in this study, the prevalence of sepsiswas approximately62%. The median (interquartile range) plasma CRP value forthe septic patients was significantly higher than that of the non-septic patients[151.5(96.4–203.0)mg/L and16.7(3.9–44.7)mg/L, respectively](p<0.01).There were also significant differences between sepsis-group and non-sepsisgroup in SOFA, FiO2/O2, temperature, mechanical ventilation, and septic shock.Multivariate logistical regression analysis showed CRP, FiO2/O2, MV weredifferent between the two groups. In the ROC curve analyses performed, thearea under the ROC curve for CRP was0.80±0.04. As a result, the best cut-offvalue for CRP to diagnose sepsis was50.7mg/L, and the sensitivity associatedwith this value was90%(76%–98%), the specificity value was83%(45%–86%).There were significant correlations between heart rate (R=0.32, p=0.01),Creatinine (R=0.36, p<0.01) and initial CRP. No differences were in initialCRP levels between survivors and non-survivors in sepsis patients. Ⅱ Procalcitonin (PCT) serum levels in terms of their value forsepsis diagnosis and prognosisA retrospective cohort study was performed at the First Hospital of JilinUniversity (Changchun, China) using data obtained from critically ill patientstreated between January1and December31,2012. Patients18years or olderwith initial PCT levels recorded in their medical histories were included in thisstudy. Patients with unclear diagnoses, and patients who were hospitalized lessthan72h, were excluded.87critically ill patients were included in this study, the prevalence of sepsiswas approximately54%. The median (interquartile range) plasma PCT value forthe septic patients was significantly higher than that of the non-septic patients10.45(1.43–63.51)ng/ml and0.37(0.07–1.92)ng/ml, respectively](p <0.01).There were also significant differences between sepsis-group and non-sepsisgroup in SOFA, FiO2/O2, tempreture and Creatinine. Multivariate logisticalregression analysis showed PCT and tempreture were different between the twogroups. In the ROC curve analyses performed, the area under the ROC curveforPCT was0.83±0.04. As a result, the best cut-off value for PCT to diagnosesepsis was1.1ng/ml, and the sensitivity associated with this value was81(70–92), the specificity value was70(56–84). There were significant correlationsbetween heart rate (R=0.33, p=0.02), Creatinine (R=0.66, p<0.01), bloodplatelet count (R=0.51, p<0.01), mean arterial pressure (R=0.49, p<0.01),septic shock (R=0.44, p<0.01), PaO2/FiO2(R=0.38, p=0.01), SOFA (R=0.67,p<0.01) and initial CRP. No differences were in initial PCT levels betweensurvivors and non-survivors in sepsis patients.Ⅲ Measuring both procalcitonin and C-reactive protein for adiagnosis of sepsis in critically ill patients A retrospective cohort study was performed at the First Hospital of JilinUniversity (Changchun, China) using data obtained from critically ill patientstreated between January1and December31,2012. Patients18years or olderwith initial PCT and CRP levels recorded in their medical histories wereincluded in this study. Patients with unclear diagnoses, and patients who werehospitalized less than72h, were excluded. The combined test was defined astesting with a single primary CRP test (or PCT test), plus PCT (or CRP) used asan additional test. A positive result of the combined test was recorded when PCTand CRP were simultaneously positive, while a negative value was recordedwhen both, or either, of the values were negative. The accuracy of the combinedtest compared with the single test was evaluated using odds ratios (ORs).55critically ill patients were included in this study and the prevalence ofsepsis was approximately60%, in the ROC curve analyses performed, the areaunder the ROC curve for PCT was0.81±0.06(95%confidence interval (CI):0.70–0.93), and for CRP was0.82±0.07(95%CI:0.67–0.96). As a result, thebest cut-off value for PCT to diagnose sepsis was1.1ng/ml, and this thresholdvalue was associated with sensitivity and specificity values of83%and68%,respectively. The best cut-off value for CRP to diagnose sepsis was50.7mg/L,and the sensitivity associated with this value was90%, the specificity value was68%, and the NPV was83%. The combination of assaying PCT and CRP levelswas associated with a sensitivity of79%, a specificity of86%, and a PPV of90%. The diagnostic OR for the combination of assaying PCT and CRP levels(19), and for levels of CRP alone (18), were also greater than that of PCT (9).However, the difference between the diagnostic ORs for the combination and forCRP alone was not statistically significant.Ⅳ C-reactive protein(CRP)and procalcitonin (PCT) in terms oftheir value of guiding antibiotic therapy in sepsis A retrospective cohort study was performed at the First Hospital of JilinUniversity (Changchun, China) using data obtained from critically ill patientstreated between January1and December31,2013. All septic patients older thanor equal to18years with discontinuing antibiotic therapy following a protocolbased on serum levels of CRP or PCT and patients were divided into2groups:PCT group and CRP group. Patients with staphylococcus aureus bacteremia,immunosuppressed patients, unclear diagnoses, and patients who werehospitalized less than72h were excluded.The cut-off for CRP was50mg/L to distinguish the severity of infection incritically ill patients, CRP less than25mg/L at D4or more than80%of CRPdecrease at D5was the threshold of stopping antibiotics. The cut-off for PCTwas1.0mg/L to distinguish the severity of infection in critically ill patients,PCTless than0.5mg/L at D4or more than80%of CRP decrease at D5was thethreshold of stopping antibiotics.113septic patients were included in this study,50patients were in CRP group and63patients were in PCT group. Nodifference in age, sex, severity scores, sites of infection and microbiology wasfound between the two groups. The common infection sites in both groups werelung and abdominal cavity. The most commonly isolated microorganisms inboth groups were Pseudomonas aeruginosa, Escherichia coli, Klebsiellapneumoniae, enterococcus.The duration of antibiotic therapy was similar in PCT group and CRP group,ICU length of stay, recurrence of the first episode of infection were notsignificant between both groups. Although patients in PCT group had the trendof the higher survival rate in the28-day mortality and survival curve analysis,the difference was not statistically significant.Ⅴ conclusionsCRP was not only associated with infection but also with many factors in septic diagnosis. So, CRP could be of value in the diagnosis of sepsis. The bestcut-off value for CRP to diagnose sepsis(50.7mg/L) was related to sensitivityand specificity of diagnosis, accuracy of which depended on precise cutoff.Increases the blood level of CRP might be a risk factor of acute kidney injury insepsis.PCT level could be as a biomarker for sepsis in critically ill patients. The bestcut-off value for PCT to diagnose sepsis was1.1ng/ml with the best sensitivityand specificity. PCT concentrations were associated with the severity of sepsisin ICU. There were positive correlations with the heart rate, blood creatinine,septic shock, SOFA and initial PCT levels, and there were negative correlationswith mean arterial pressure, blood platelet, PaO2/FiO2and initial PCT levels.PCT levels was associated with sepsis-related AKI.For the critically ill patients examined in this cohort, CRP was found to besuitable for eliminating a diagnosis of infection with high negative predictivevalue; PCT was not found to be a good marker for diagnosing sepsis. However,the use of a ‘both positive’ format for the assays of PCT and CRP levels wasfound to be suitable for obtaining a deterministic diagnosis of infection withgreater specificity and positive predictive value.C-reactive protein was as useful as procalcitonin in reducing antibiotic use inseptic patients.Ⅵ The novel findings in this thesisThe best cut-off value for CRP and PCT to diagnose sepsis was50.7mg/L,1.1ng/ml respectively. This study evaluated the usefulness of assaying PCT andCRP for the identification of sepsis in critically ill adult patients. CRP was foundto be suitable for eliminating a diagnosis of infection due to its greater negativepredictive value. In contrast, PCT was not found to be a good marker for diagnosing sepsis. However, the use of a ’both positive’ format for the assays ofPCT and CRP levels was found to be suitable for obtaining a deterministicdiagnosis of infection due to the specificity and positive predictive valueassociated with this combination of biomarkers. This study established aflowchart based on CRP for the decision to discontinue antibiotic treatment andproved its practicability, and verified that CRP was as useful as PCT in reducingantibiotic use in septic patients. |
PDF Full Text Request