Cutoff values for INR, platelet count, and fibrinogen were determined in the same manner using the median.
Data are presented as means and SD as well as median and 25th and 75th percentiles. The Shapiro-Wilk excluded normality of some data sets (data not shown). Therefore, the Mann-Whitney U test was used for comparison of means. Dependence of 30-day survival on thromboelastometry score was evaluated using Kaplan-Meier analysis and the log-rank test. Furthermore, odds ratio, sensitivity, and specificity were determined, as appropriate. For multivariate analysis of possible predictors of 30-day mortality, binary logistic regression was used. SPSS version 16 (SPSS Inc, Chicago, Ill) was used for all statistical procedures. An a priori α error P < 0.05 was considered to indicate statistical significance.
Distribution of thromboelastometry variables in patients with sepsis
The thromboelastometry values obtained from all 98 septic patients showed a broad distribution. Mean values and SD (median; 25th percentile, 75th percentile) of thromboelastometry variables were 534 ± 213 s (482, 406, 616 s) for CT, 235 ± 162 s (185, 127, 291 s) for CFT, 55.0 ± 12.0 mm (55, 47, 64 mm) for MCF, and 56.2 ± 12.5 degrees (58, 47, 67 degrees) for α angle.
Thromboelastometry variables in survivors and nonsurvivors of sepsis
Values for CFT, MCF, and α angle were different between survivors and nonsurvivors. In nonsurvivors, CFT was prolonged (276 ± 193 s vs. 194 ± 109 s, P = 0.021), MCF was reduced (52.0 ± 12.1 mm vs. 57.3 ± 11.5 mm, P = 0.042), and the α angle was less (53.4 ± 12.8 degrees vs. 58.9 ± 11.8 degrees, P = 0.028). Clotting time was not significantly different between survivors and nonsurvivors.
Kaplan-Meier curves of individual thromboelastometry variables
Clotting time, CFT, MCF, and α angle were classified using the median of the cohort as the discriminating value (see Materials and Methods). Kaplan-Meier curves for 30-day survival in patients with and without pathological thromboelastometry values (CT, CFT, MCF, and α angle) reveal that CFT, MCF, and α angle are associated with outcome (Fig. 1).
Survival of patients with sepsis with and without pathological thromboelastometry findings
In 35 patients, CFT, MCF, and α angle were normal, whereas at least one of these variables was pathological in 63 patients. Kaplan-Meier analysis revealed that the prognosis was much better in the patients without pathological thromboelastometry findings (Fig. 2). Here, 30 of 35 patients survived for at least 30 days (85.7%). In contrast, only 37 (58.7%) of 63 patients survived 30 days if there was at least one pathological thromboelastometry finding (P = 0.005, log-rank test). Determination of the odds ratio revealed a 4.1-fold increased 30-day mortality in patients when there was at least one pathological thromboelastometry finding; sensitivity and specificity were 83.8% and 44.8%, respectively.
Comparison of thromboelastometry and SAPS II and SOFA scores
Whereas thromboelastometry variables CFT, MCF, and α angle were different in survivors and nonsurvivors, both SAPS II (54.1 ± 13.7 vs. 48.4 ± 16.0; P = 0.069) and SOFA scores (13.2 ± 3.8 vs. 11.7 ± 3.7; P = 0.066) showed no difference. Thus, univariate analysis suggests that thromboelastometry might better predict outcome than SAPS II and SOFA scores.
Multivariate analysis using logistic regression analysis, which included thromboelastometry (patients with and without pathological CFT, MCF, and α angle, respectively) as well as SAPS II and SOFA scores, confirmed that thromboelastometry is an important and independent prognostic factor for 30-day survival (odds ratio, 4.1; 95% confidence interval, 1.393-11.923; P = 0.01), whereas SAPS II and SOFA scores were excluded. To exclude that colinearity affected the logistic regression analysis, we repeated analysis with thromboelastometry and either SAPS II score or SOFA score; this procedure resulted in almost identical results.
Comparison of thromboelastometry and INR, fibrinogen, and platelets
International normalized ratio was different in survivors and nonsurvivors (1.27 ± 0.46 vs. 1.53 ± 0.48; P = 0.001) (Table 1). Defining the cutoff of the INR as the median of all patients with sepsis (see Materials and Methods), INR was associated with a survival of 81.2% (INR <1.27) and 56.1% (INR ≥1.27), respectively. Determination of the odds ratio revealed a 3.4-fold increased 30-day mortality in patients with prolonged INR. Sensitivity and specificity of INR for the prediction of mortality in sepsis were 66.6% and 62.9%, respectively.
Moreover, platelet count was markedly higher in survivors than in nonsurvivors (161.2 ± 108.1 vs. 117.6 ± 106.5; P = 0.023). Kaplan-Meier statistics demonstrated a 30-day survival of 80.4% and 57.4%, respectively, when the median of platelet count was chosen as the cutoff. The odds ratio was 3.0; sensitivity and specificity were 68.9% and 42.2%, respectively.
Fibrinogen values showed no significant differences in survivors and nonsurvivors of sepsis in the present study. Multivariate analysis (logistic regression analysis) including INR, fibrinogen, platelet count, and thromboelastometry (patients with and without pathological findings, respectively) demonstrated that thromboelastometry is an important and independent prognostic factor for 30-day survival (odds ratio, 5.6; confidence interval, 1.49-20.7; P = 0.01), while INR, fibrinogen, and platelet count were excluded. Again, colinearity was excluded as the reason for this result.
Our results demonstrate for the first time that thromboelastometry detects differences in hemostasis between survivors and nonsurvivors of severe sepsis and that the absence and presence of pathological thromboelastometry findings predict 30-day survival better than the SAPS II and SOFA scores.
Thromboelastography, first described by Hartert in 1948 (16), and rotational thromboelastometry are point-of-care methods measuring the kinetics of clot formation in whole-blood samples by describing the viscoelastic properties of the forming clot (16, 17). Thromboelastometry is a widely accepted method in cardiac and liver transplantation surgery (18, 19). The values of variables obtained by this method include the CT, representing the time interval to onset of coagulation, the CFT and α angle, describing the kinetics of clot formation, and the maximal clot firmness, which depends on both platelet count and fibrin polymerization. In our assay, any effects of heparin administered to the patients and causing potential differences in clot formation were abolished by adding heparinase to the assays in sufficient concentrations.
Recently, results obtained from animal experiments and in a small patient cohort have suggested that sepsis-evoked alterations in hemostasis might be detected by thromboelastometry (10-12). It is unknown, however, whether thromboelastometry variables might predict outcome of sepsis.
In the present study, prognosis of septic patients was markedly different in patients with and without coagulopathy, and 30-day survival decreased from 85.7% to 58.7%, if thromboelastometry revealed alterations of hemostasis. The changes in thromboelastometry variables, which are associated with increased mortality, most likely are indicative for the depletion of coagulation factors and platelets. As the reason for derangement of hemostasis, DIC can be presumed, although we have not assessed this issue in detail using further laboratory tests. However, this notion is in accordance with the fact that DIC, which can evoke organ ischemia and multiorgan failure, is associated with increased mortality in sepsis (3).
One established method to detect DIC is a sophisticated scoring system encompassing platelet count, INR, fibrinogen, and a fibrinolysis marker. Using such a complex scoring system, groups with low (15%) and high mortality (40%) have been discriminated (1). Remarkably, however, the results of our study demonstrate that the presence or absence of a single pathological thromboelastometry finding is associated with marked differences in survival and allows discriminating two patient subsets with a 58.7% and 85.7% 30-day survival, respectively. Furthermore, comparison of thromboelastometry with INR and platelet count in the present study demonstrated that thromboelastometry allowed the best prediction of outcome.
It is an important finding of the present study that both univariate and multivariate analyses demonstrated the importance of thromboelastometry in comparison to the SAPS II and SOFA scores. The fact that thromboelastometry proved to be both an important and independent predictor of 30-day survival, whereas the SAPS II and SOFA scores were not, strongly suggests that the coagulation system is of outstanding importance for the prognosis in severe sepsis, obviously to a greater extent than the abnormalities reflected by the SAPS II and SOFA scores. In this regard, it is also important to note that modulation of hemostasis in sepsis improves the prognosis in both experimental settings and potentially in patients (5, 6).
Our finding that early thromboelastometry can detect patients with a worse prognosis suggests that this method may be used to identify patients in which therapeutic modulation of hemostasis might be promising. Furthermore, thromboelastometry might also be capable to monitor the efficacy of such pharmacological interventions. Both issues are important because thromboelastometric results are readily available at the bedside, and an early goal-directed therapy in sepsis is mandatory to improve the prognosis (20).
It is important to state that the study has limitations: cutoff values and survival were determined in the same collective. It is therefore important to confirm the diagnostic values of the cutoff values of the present study in a second data set. Whereas a moderate correlation of DIC score and thromboelastometry has been demonstrated in a recent study, a direct comparison of the association of thromboelastometry and DIC score with outcome is desirable in a future study (12).
The results of the present study demonstrate that thromboelastometry is a better predictor of 30-day survival in severe sepsis than the SAPS II and SOFA scores. These results underline the pathophysiological importance of hemostasis and the relevance of whole-blood coagulation tests. Further studies are necessary to investigate whether thromboelastometry-guided treatments can improve the prognosis of sepsis.
1. Bakhtiari K, Meijers JC, de Jonge E, Levi M: Prospective validation of the International Society of Thrombosis and Haemostasis scoring system for disseminated intravascular coagulation. Crit Care Med
2. Lissalde-Lavigne G, Combescure C, Muller L, Bengler C, Raillard A, Lefrant JY, Gris JC: Simple coagulation tests improve survival prediction in patients with septic shock. J Thromb Haemost
3. Franchini M, Lippi G, Manzato F: Recent acquisitions in the pathophysiology, diagnosis and treatment of disseminated intravascular coagulation. Thromb J
4. Amaral A, Opal SM, Vincent JL: Coagulation in sepsis. Intensive Care Med
5. Dhainaut JF, Yan SB, Joyce DE, Pettilä V, Basson B, Brandt JT, Sundin DP, Levi M: Treatment effects of drotrecogin alfa (activated) in patients with severe sepsis with or without overt disseminated intravascular coagulation. J Thromb Haemost
6. Levi M, van der Poll T: Coagulation in sepsis: all bugs bite equally. Crit Care
7. Baratto F, Michielan F, Meroni M, Dal Palù A, Boscolo A, Ori C: Protein C concentrate to restore physiological values in adult septic patients. Intensive Care Med
8. Bernard GR, Vincent JL, Laterre PF, LaRosa SP, Dhainaut JF, Lopez-Rodriguez A, Steingrub JS, Garber GE, Helterbrand JD, Ely EW, et al: Recombinant Human Protein C Worldwide Evaluation in Severe Sepsis (PROWESS) study group. Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med
9. Ganter MT, Hofer CK: Coagulation monitoring: current techniques and clinical use of viscoelastic point-of-care coagulation devices. Anesth Analg
10. Zacharowski K, Sucker C, Zacharowski P, Hartmann M: Thrombelastography for the monitoring of lipopolysaccharide induced activation of coagulation. Thromb Haemost
11. Velik-Salchner C, Streif W, Innerhofer P, Maier S, Knotzer H, Pajk W, Klingler A, Mittermayr M, Haas T: Endotoxinemia-induced changes in coagulation as measured by rotation thrombelastometry technique and conventional laboratory tests: results of a pilot study on pigs. Blood Coagul Fibrinolysis
12. Sivula M, Pettilä V, Niemi TT, Varpula M, Kuitunen AH: Thromboelastometry in patients with severe sepsis and disseminated intravascular coagulation. Blood Coagul Fibrinolysis
13. Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D, Cohen J, Opal SM, Vincent JL, Ramsay G, et al: 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med
14. Le G Jr, Lemeshow S, Saulnier F: A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA
15. Ferreira FL, Bota DP, Bross A, Melot C, Vincent JL: Serial evaluation of the SOFA score to predict outcome in critically ill patients. JAMA
16. Hartert H: Blutgerinnungsstudien mit der Thrombelastographie, einem neuen Untersuchungsverfahren. Klin Wschr
17. Calatzis A, Fritzsche P, Calatzis A, Kling M, Hipp R, Stemberger A: roTEG Coagulation Analyzer-Ein neues System zur intraoperativen Bedside-Gerinnungsdiagnostik. Anaesthesist
18. Gillies BS: Thromboelastography and liver transplantation. Semin Thromb Hemost
21(suppl 4):45-49, 1995.
19. Dunning J, Versteegh M, Fabbri A, Pavie A, Kolh P, Lockowandt U, Nashef SA: EACTS Audit and Guidelines Committee. Guideline on antiplatelet and anticoagulation management in cardiac surgery. Eur J Cardiothorac Surg
20. Lin SM, Huang CD, Lin HC, Liu CY, Wang CH, Kuo HP: A modified goal-directed protocol improves clinical outcomes in intensive care unit patients with septic shock: a randomized controlled trial. Shock
Sepsis; thromboelastometry; survival; SAPS II score; SOFA score