BJA Advance Access originally published online on July 18, 2006
British Journal of Anaesthesia 2006 97(4):503-508; doi:10.1093/bja/ael181
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Procalcitonin for early prediction of survival outcome in postoperative critically ill patients with severe sepsis
1 Department of Anaesthesiology and Intensive Care Medicine, Graz Medical University Graz, Austria
2 Biomedical Engineering and Computing Unit of the Department of Surgery, Graz Medical University Graz, Austria
*Corresponding author. E-mail: ashraf.dahaba{at}meduni-graz.at
Accepted for publication June 14, 2006.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and methodsResultsDiscussionReferences |
|---|
Background. Identification of postoperative patients at high risk of dying early after intensive care unit (ICU) admission through a fast and readily available parameter may help in determining therapeutic interventions or further diagnostic procedures that could have an impact on patients' outcome. The aim of our study was to assess the utility of procalcitonin (PCT) and other readily available parameters, as useful early (days 1–3) predictors of mortality in postoperative patients diagnosed with severe sepsis within 24 h preceding their operation.
Methods. More than a period of 2 yr, subsets of 69 postoperative patients admitted with severe sepsis and 890 non-septic ICU patients were investigated. PCT, C-reactive protein (CRP) and sequential organ failure assessment (SOFA) score were recorded over the duration of ICU stay.
Results. PCT area under receiver operating characteristic (ROC) curve was 0.78 on day 3 and was highly predictive of fatal outcome (0.90) at day 6. Area under ROC curve of SOFA score was 0.85 on day 3 and remained in this range until day 6. Area under ROC curves on day 3 of CRP (0.61) was non-predictive and remained non-predictive over the duration of ICU stay.
Conclusions. PCT exhibited no discriminative power early after ICU admission for prediction of mortality in critically ill patients with severe sepsis, compared with a high predictive power of SOFA score on day 3. However, using PCT could still serve as a useful complementary comparator for prediction of survival outcome using the SOFA score.
Keywords: complications, sepsis; protein, C-reactive; statistics, receiver operating characteristic; survival, outcome
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Sepsis is the systemic inflammatory response to microbial infection.1 Normal immune function depends on an appropriate mediator response that often prevents the progression of infection. However, an exaggerated immune response, with overproduction of inflammatory mediators, is the pivotal mechanism in the sequence of events that results in diffuse injury of healthy tissues, major organs dysfunction and the associated mortality.2
Despite an extending knowledge of mediators and the mechanisms involved in systemic inflammation, sepsis is still one of the major causes of death in critically ill patients.3 Hence, there is a growing need for precise staging and prognostication tools for patients with severe sepsis. Identification of patients at high risk of dying, early after intensive care unit (ICU) admission, through a fast and readily available laboratory parameter may help in determining therapeutic interventions, such as changes in therapeutic protocols or further diagnostic procedures aiming at preventing shock and multiple organ failure with all their sequels that could have an impact on patients' outcome.4
During bacterial infection the procalcitonin (PCT) CT-messenger RNA CALC-I gene ion is ubiquitously expressed from various extrathyroid neuroendocrine tissues throughout the body.5 Hence, the term ‘hormokine’ was coined to signify this cytokine-like host-response to sepsis by a peptide of an otherwise classical endocrine family.6 Several studies found that in critically ill patients with sepsis a steady increase in PCT concentrations would indicate poor outcome.7–11 However, to date no study has precisely evaluated the discriminative power of PCT as an early after ICU admission predictor of survival outcome in non-cardiac surgery patients admitted to the ICU with severe sepsis. The aim of our study was to assess the utility of PCT, C-reactive protein (CRP) and sequential organ failure assessment (SOFA) score12 as useful predictors of mortality early after ICU admission (days 1–3) in patients with severe sepsis.
|
Patients and methods |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Our study was designed according to Randolph AG criteria,13 namely it was conducted in a representative sample of patients that was sufficiently homogenous with respect to prognostic outcome. Thus, we only included patients according to the recently revisited sepsis consensus definition ‘PIRO’,1 namely patients admitted to surgical ICU after potentially septic operations who were first diagnosed with severe sepsis within 24 h preceding their operation. This eliminated the confounding factor of a variable onset of severe sepsis.14 The unbiased and well-defined endpoint used in our study was mortality directly related to severe sepsis within the first 28 days of ICU stay.13 Thus, patients who developed severe sepsis after ICU admission and patients with severe sepsis who died from underlying causes not related to severe sepsis such as massive myocardial infarction or direct damage to internal organs as a result of multiple trauma were all excluded from the outcome analysis.
More than a period of 2 yr, subsets of 69 patients admitted to the university hospital surgical ICU with severe sepsis and 890 non-septic postoperative ICU patients with normal PCT values were investigated. Severe sepsis patients had microbiologically proven infectious aetiology arising from urinary, pulmonary, intra-abdominal or bloodstream infections. In addition, patients had clinical evidence of sepsis, according to the recently approved International Sepsis Consensus Conference definitions,1 namely body temperature of <36 or >38°C, heart rate >90 beats min–1, leucocyte count of <4000 or >12 000 cells µl–1, tachypnea of >20 bpm or hyperventilation as indicated by an arterial partial pressure of carbon dioxide of <32 mm Hg. Severe sepsis was defined as sepsis associated with hypotension or hypoperfusion manifesting as lactic acidosis, oliguria or acute alteration in mental state.1
All patients received culture-guided antibiotic treatment and haemodynamic support therapy. PCT and CRP were recorded at admission and over the duration of ICU stay. PCT was assayed using immunoluminometric technique (LUMItestTM, B·R·A·H·M·S Diagnostica, Berlin, Germany). CRP was assayed using commercially available kits.
Statistical analysis
Based upon a previous study in ICU patients who developed sepsis after cardiac surgery with cardiopulmonary bypass,9 which showed that with PCT area under receiver operating characteristic (ROC) curve of 0.87 and 84/88 positive/negative predictive values (PPV/NPV) at day 3, the difference between the median (quartiles) PCT values of 1.8 (0.4–4.4) ng ml–1 in survivors and 16 (6.1–32.9) ng ml–1 in non-survivors was 14.2 ng ml–1. After estimation of standard deviations and logarithmic transformations to accommodate for non-linear data distribution, our a priori power analysis for two-sided t-test showed that a sample size of 15 non-survivors would be required to reveal a statistically significant difference between survivors and non-survivors with 90% power.
For survival outcome specificity, sensitivity, PPV and NPV were calculated. The likelihood ratio (LHR) was calculated as sensitivity/(1–specificity). Area under the ROC curves were constructed according to Hanley and McNeil by plotting the sensitivity against 1–specificity.15 An area under ROC curve of 1 indicates a perfect predictive power, and the closer the area under the ROC curve to 1 the greater the discriminative power of the marker. Whereas, a value of 0.5 indicates a non-informative prediction that predicts fatal outcome no better than a coin toss. Area under ROC curve >0.8 was considered good prediction. Data were expressed as area under ROC curve and 95% confidence intervals (CI). We further compared the area under ROC curves from different parameters using the Hanley and McNeil method.16 One way analysis of variance (ANOVA) was used for the analysis of differences between survivors and non-survivors. Data were expressed as mean (SD). P<0.05 was considered statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Patients' patient characteristics, reason for ICU admission and individual isolates are presented in Tables 1–3. Within the study period, out of the 69 severe sepsis patients, 18 patients died. Whereas, out of the 890 non-septic patients, with normal PCT plasma concentrations, 49 patients died. This yielded a mortality odds ratio of 4.7, CI 2.3–7.6.
|
At admission there was no difference in the mean PCT plasma concentrations between survivors and non-survivors. However, after admission, mean PCT plasma concentration in survivors declined towards the normal range compared with a steady increase in non-survivors, with a significant difference (P<0.05) between survivors and non-survivors starting from day 4.
There was no significant difference in the SOFA score at admission; however, during the ICU treatment period the mean SOFA score was significantly higher (P<0.01) in non-survivors [16.7 (8.1)] compared with survivors [3.1 (1.8)] (Fig. 1).
|
Area under ROC curve of PCT at admission was not predictive of outcome. PCT area under ROC steadily increased to 0.78 on day 3. PCT was highly predictive (0.9) at day 6 and remained in this range until day 9 (Figs 2 and 3). PCT cut-off value of 3.2 ng ml–1 at day 6 was associated with the optimal combination of sensitivity (0.85), specificity (0.89), PPV/NPV (0.77/0.96) and LHR (7.73). A similar combination was found at day 7 with a lower PCT cut-off value of 2.5 ng ml–1, 1.7 ng ml–1 at day 8 and 1.4 ng ml–1 at day 9. Area under ROC curve of SOFA score on day 3 was 0.85, CI 0.78–0.94 and remained in this range until day 6. The correlation coefficient between the PCT and SOFA score area under ROC curves on day 3 was 0.73.
|
|
When PCT values were plotted backwards to end with the day the patients died (Fig. 4), it became clear that after the PCT peak, PCT significantly declined before the patients died (P=0.0001).
|
During the ICU treatment period, there was no significant difference in the mean CRP between survivors [18.9 (24.1) mg dl–1] and non-survivors [22.9 (11.3) mg dl–1]. The area under ROC curves for CRP (0.61) on day 3 did not increase over the following days.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Our study revealed that the predictive power of PCT peaked at day 6. This seems to be inferior to the predictive value of clinical assessment with SOFA score, which predicted lethal outcome as early as day 3 for a mean length of 12 days ICU stay. Because the development of multiple organ failure is the main cause of mortality in patients with severe sepsis, assessing morbidity using an organ dysfunction score such as the SOFA score seems to have a better predictive value than laboratory parameters. Furthermore, as a result of the fact that at the day PCT gives a significant mortality predictive value (day 4 and later), the majority (56%) of the non-survivors will die just a few days later (between days 7 and 12). These patients dying from septic multiorgan failure, as indicated by their highly elevated PCT concentrations of >45 ng ml–1 (Fig. 4) and SOFA score of >18 (Fig. 1) were probably easy to identify as being in an acute life-threatening condition by the deterioration of their laboratory and clinical parameters. On the other hand, using the Hanley and McNeil method16 the predictive value of PCT on day 3 correlated to the predictive value of the SOFA score. This is in accordance with a recent report of PCT concentrations highly correlating with increasing categories of SOFA score in septic patients.17 Thus, using a laboratory sepsis marker such as PCT could still serve as a useful complementary comparator to the SOFA score in assessing the severity of morbidity and predicting survival outcome.
In our study population, patients with severe sepsis were at almost five times higher risk of dying than non-septic patients. Early identification of patients at risk of dying was reported to have a major impact on survival outcome as the 16% mortality incidence in septic patients increased to 20% mortality with severe sepsis and further increased to 46% mortality in patients who proceeded to septic shock.4 In this regard PCT represents a valuable prognostic marker as our results revealed that statistically speaking if a patient was to survive severe sepsis, PCT had to decline below the cut-off value of 3.2 ng ml–1 at day 6, 2.5 ng ml–1 by day 7, 1.7 ng ml–1 by day 8 and below 1.4 ng ml–1 by day 9, a clear indication of the rapid progression of the sepsis cascade in patients with lethal outcome.
When we examined the last ICU days of patients with lethal outcome (Fig. 4), it became clear that after PCT peak, PCT significantly declined before death, a phenomenon that could confound, at the late stages of sepsis, any therapeutic strategy aiming at lowering PCT. This peculiar phenomenon was previously reported in severely burned septic patients18 and in patients with septic shock.19 In animal studies, injection of the ‘proximal’ proinflammatory mediator tumour necrosis factor-
(TNF-) induced a 25-fold massive and sustained elevation of PCT,20 clearly indicating that PCT is an ‘intermediary’ mediator in the cascade of events contributing to a lethal outcome. Thus, a decline in PCT before death may involve either a failure in PCT production per se or failure in another critical reactant or synergism in the complicated sepsis cascade. We can only speculate that patients could have manifested a deficiency in their ability to mount an effective defensive mechanism before death, in what seems to be a very ominous sign of a poor outcome.
The lack of CRP discriminative power that we demonstrated in our study is because of the fact that serum CRP, an acute phase protein widely used to support the diagnosis of infection, is not exclusively specific for sepsis as it increases in various conditions associated with tissue injury.21
In conclusion, PCT exhibited no discriminative power early after ICU admission for prediction of mortality in critically ill patients with severe sepsis, compared with a high predictive power of SOFA score on day 3. However, using PCT could still serve as a useful complementary comparator for prediction of survival outcome using the SOFA score.
|
|
|
Acknowledgments |
|---|
The authors would like to thank Dr Beatrix Kirsten (B·R·A·H·M·S Diagnostica, Austria) for her great help in procalcitonin data collection, data tabulation and data preparation. Her great work and meticulous efforts were indeed a valuable contribution to the study.
|
References |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
1 Levy MM, Fink MP, Marshall JC, et al. for members of the Society of Critical Care Medicine (SCCM)/European Society of Intensive Care medicine (ESICM)/American College of Chest Physicians (ACCP)/American Thoracic Society (ATS)/Surgical Infection Society (SIS). 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med 2003; 31:1250–6[CrossRef][Web of Science][Medline]
2 Bone RC. The pathogenesis of sepsis. Ann Intern Med 1991; 115:457–69
3 Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Garcillo J, Pinsky MR. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med 2001; 29:1303–10[CrossRef][Web of Science][Medline]
4 Rangel-Frausto MS, Pittet D, Costigan M, Hwang T, Davis CS, Wenzel RP. The natural history of the systemic inflammatory response syndrome (SIRS). A prospective study. JAMA 1995; 273:117–23
5 Mueller B, White JC, Nylen ES, Snider RH, Becker KL, Habener JF. Ubiquitous expression of the calcitonin-1 gene in multiple tissues in response to sepsis. J Clin Endocrinol Metab 2001; 86:396–404
6 Becker KL, Nylen ES, Snider RH, Mueller B, White JC. Immunoneutralization of procalcitonin as therapy of sepsis. J Endotoxin Res 2003; 9:367–74[CrossRef][Web of Science]
7 Meisner M, Tschaikowsky K, Palmaers T, Schmidt J. Comparison of procalcitonin (PCT) and C-reactive protein (CRP) plasma concentrations at different SOFA scores during the course of sepsis and MODS. Crit Care 1999; 3:45–50[CrossRef][Web of Science][Medline]
8 Herrmann W, Ecker D, Quast S, Klieden M, Rose S, Marzi I. Comparison of procalcitonin, sCD14 and interleukin-6 values in septic patients. Clin Chem Lab Med 2000; 38:41–6[CrossRef][Web of Science][Medline]
9 Adamik B, Kuebler-Kielb J, Golebiowska B, Gamian A, Kuebler A. Effect of sepsis and cardiac surgery with cardiopulmonary bypass on plasma level of nitric oxide metabolites, neopterin, and procalcitonin: correlation with mortality and postoperative complications. Intensive Care Med 2000; 26:1259–67[CrossRef][Web of Science][Medline]
10 Tschaikowsky K, Hedwig-Geissing M, Schiele A, Bremer F, Schywalsky M, Schuettler J. Coincidence of pro- and anti-inflammatory responses in the early phase of severe sepsis: longitudinal study of mononuclear histocompatibility leukocyte antigen-DR expression, procalcitonin, C-reactive protein, and changes in T-cell subsets in septic and postoperative patients. Crit Care Med 2002; 30:1015–23[CrossRef][Web of Science][Medline]
11 Wunder C, Eichelbroenner O, Roewer N. Are IL-6, IL-10 and PCT plasma concentrations reliable for outcome prediction in severe sepsis? A comparison with APACHE III and SAPS II. Inflamm Res 2004; 53:158–63[CrossRef][Web of Science][Medline]
12 Vincent JL, De Mendonca A, Cantraine F, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Crit Care Med 1998; 26:1793–800[Web of Science][Medline]
13 Randolph AG, Guyatt GH, Richardson WS. Prognosis in the intensive care unit: finding accurate and useful estimates for counseling patients. Crit Care Med 1998; 26:767–72[CrossRef][Web of Science][Medline]
14 Teplick R. A predilection for posterior prediction and phenotypic precision. Crit Care Med 2002; 30:481–3[CrossRef][Web of Science][Medline]
15 Hanley JA and McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982; 143:29–36
16 Hanley JA and McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983; 148:839–43
17 Castelli GP, Pognani C, Meisner M, Stuani A, Bellomi D, Sgarbi L. Procalcitonin and C-reactive protein during systemic inflammatory response syndrome, sepsis and organ dysfunction. Crit Care 2004; 8:R234–42[CrossRef][Web of Science][Medline]
18 Von Heimburg D, Stieghorst W, Khorram-Sefat R, Pallua N. Procalcitonin-a sepsis parameter in severe burn injuries. Burns 1998; 24:745–50[CrossRef][Web of Science][Medline]
19 Schroeder J, Staubach KH, Zabel P, Stueber F, Kremer B. Procalcitonin as a marker of severity in septic shock. Langenbecks Arch Surg 1999; 384:33–8[CrossRef][Web of Science][Medline]
20 Whang KT, Vath SD, Nylen ES, et al. Procalcitonin and proinflammatory cytokine interactions in sepsis. Shock 1999; 12:268–73[Web of Science][Medline]
21 Pepys MB and Baltz ML. Acute phase proteins with special reference to C-reactive protein and related proteins (pentaxins) and serum amyloid A protein. Adv Immunol 1983; 34:141–212[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. Zhang, T. Lavaux, A.-C. Voegeli, T. Lavigne, V. Castelain, N. Meyer, R. Sapin, D. Aunis, M.-H. Metz-Boutigue, and F. Schneider Prognostic Value of Chromogranin A at Admission in Critically Ill Patients: A Cohort Study in a Medical Intensive Care Unit Clin. Chem., September 1, 2008; 54(9): 1497 - 1503. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. L. Carboni, R. Fahrner, A. Gazdhar, G. Printzen, R. A. Schmid, and B. Hoksch Comparison of procalcitonin and CrP in the postoperative course after lung decortication Eur. J. Cardiothorac. Surg., May 1, 2008; 33(5): 777 - 780. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||