Day 3 versus day 1 disseminated intravascular coagulation score among sepsis patients

Disseminated intravascular coagulation in sepsis patients

57

Anaesth Intensive Care 2016 | 44:1

Day 3 versus Day 1 disseminated intravascular coagulation score among sepsis patients: a prospective observational studyJ. Y. Park*, S. Park†, S. Y. Park‡, Y. S. Sim§, J. H. Kim**, Y. I. Hwang††, S. H. Jang††, K. S. Jung‡‡

SummaryThe role of disseminated intravascular coagulation (DIC) has not been extensively studied in patients with sepsis. A prospective study was performed in a single university hospital. The incidences of DIC at day 1 (<24 hours post-sepsis diagnosis) and day 3 (48 to 72 hours) were investigated among patients with sepsis. The International Society of Thrombosis and Haemostasis criteria for DIC were used. Among 381 patients initially screened, 219 were enrolled in this study and the incidences of overt DIC were 27.9% and 30.1% on day 1 and day 3, respectively. Patients with pneumonia had a lower incidence of DIC on day 1, but a higher hospital mortality rate compared to those with non-pneumonia sepsis. In multivariate models, although day 1 and day 3 DIC scores were not associated with hospital mortality after adjusting for existing severity scores, the change in DIC scores (odds ratio 1.862; 95% confidence interval 1.061 to 3.266) exhibited a significant association. Day 3 DIC scores were more accurate in predicting hospital mortality than day 1 DIC scores (P <0.001), especially in patients with non-pneumonia sepsis. However, DIC scores did not give additional discriminative power to the existing prognostic scores in predicting mortality of patients with sepsis. In conclusion, the change in DIC score was significantly associated with hospital mortality. Patients with pneumonia sepsis had a lower incidence of DIC on day 1, despite their higher disease severity and mortality rate, compared to those with other sources of sepsis.

Key Words: disseminated intravascular coagulation, mortality, sepsis

Sepsis is a major cause of death worldwide, and an excessive host response in sepsis plays an important role in the occurrence of disseminated intravascular coagulation (DIC). Previously, DIC was known to arise in 35% of patients with severe sepsis and its occurrence is associated with worsening organ failure1–3. Although the understanding and management of sepsis has markedly improved over the past decades, effective treatments targeting DIC per se, are very limited4.

There has been an increased number of studies providing new insight into the role of coagulation activation in sepsis5,6.

Specifically, data from a well-controlled clinical trial, the Recombinant Human Activated Protein C Worldwide Evaluation in Severe Sepsis (PROWESS) study, demonstrated that abnormalities in biomarkers of coagulopathy, inflammation, and endothelial injury were apparent in patients with severe sepsis7. In surgical patients, following the first 48 hours of ICU admission, platelet counts tended to improve in survivors, but remained unchanged or worsened in non-survivors8. Similarly, other authors have also emphasised the importance of the dynamic evolution of coagulation markers because sepsis-associated coagulopathy can be subtle9,10.

Without appropriate treatment during the early period, patients with sepsis can develop hypodynamic shock when tissue hypoperfusion and coagulopathy in the microcirculation progress (i.e. DIC)11. In addition, some experts suggest there is a period of time in which coagulation activation leads to a detrimental process (i.e. tissue damage)6. Therefore, in this study, we hypothesised that the occurrence of DIC at day 3 (48 to 72 hours post-sepsis diagnosis) would reveal prognostic information different from DIC at day 1, and that the incidence would also vary according to the origin of sepsis. To diagnose DIC, we used the International Society of Thrombosis and Haemostasis (ISTH) criteria, which have been prospectively validated1,12. We investigated the impact of DIC scores (i.e. day 1, day 3, and the change between DIC scores [∆DIC]) on hospital mortality; we also compared these results with respect to different origins of sepsis.

* MD, Pulmonologist, Division of Pulmonary, Allergy and Critical Care Medicine, Department of Internal Medicine, Hallym University Sacred Heart Hospital, Seoul, Korea

† MD PhD, Intensivist and Pulmonologist, Division of Pulmonary, Allergy and Critical Care Medicine and Department of Internal Medicine, Hallym University Sacred Heart Hospital, Seoul, Korea

‡ MD, Pulmonologist, Division of Pulmonary, Allergy and Critical Care Medicine, Department of Internal Medicine, Hallym University Sacred Heart Hospital, Seoul, Korea

§ MD, Intensivist and Pulmonologist, Division of Pulmonary, Allergy and Critical Care Medicine and Department of Internal Medicine, Kangnam Sacred Heart Hospital, Seoul, Korea

** MD, Allergist and Pulmonologist, Division of Pulmonary, Allergy and Critical Care Medicine, Department of Internal Medicine, Hallym University Sacred Heart Hospital, Seoul, Korea

†† MD PhD, Intensivist and Pulmonologist, Division of Pulmonary, Allergy and Critical Care Medicine and Department of Internal Medicine, Hallym University Sacred Heart Hospital, Seoul, Korea

‡‡ MD PhD, Pulmonologist, Division of Pulmonary, Allergy and Critical Care Medicine, Department of Internal Medicine, Hallym University Sacred Heart Hospital, Seoul, Korea

Address for correspondence: Dr Sunghoon Park. Email: [email protected] for publication on October 20, 2015

J. Y. Park et al

58


Materials and methods

Study populationThis prospective study was performed in a tertiary

academic hospital (830 beds) from 1 August, 2010 to 31 July, 2012. Adult patients aged 18 years and older who were admitted to the medical ICU for sepsis (or severe sepsis or septic shock) were initially screened. Exclusion criteria were patients who had been receiving anticoagulation treatment, patients who died within 72 hours of sepsis diagnosis, patients with liver cirrhosis (≥ Child-Pugh B), cardiopulmonary resuscitation at ICU admission, refusal to consent, do-not-resuscitate status, transfer to other hospital, uncontrolled haematologic malignancy or solid cancer, and use of steroid or immunosuppressant agents. However, patients were eligible if their cancers had been in complete remission for more than six months or if they were receiving a low-dose steroid (i.e., a dose of ≤10 mg/day prednisolone or equivalents). For the diagnosis of sepsis, the 2008 Surviving Sepsis Campaign guidelines were used. Sepsis was defined as the probable or documented presence of infection, together with systemic manifestations of infection; severe sepsis was defined as sepsis plus sepsis-induced organ dysfunction or tissue hypoperfusion; septic shock was defined as severe sepsis with hypotension despite adequate fluid resuscitation13. Community-acquired infections were defined as infections that occurred in the community or in hospital during the incubation period (<48 hours after hospitalisation). However, it was classified as a healthcare-associated infection if a patient met any of the healthcare-risk factors as described in a report by Friedman et al14. Hospital-acquired infections were defined as infections that occurred >48 hours after hospitalisation. This study was approved by the Hallym University Institutional Review Board (Approval No.: 2010-I060). We obtained written, informed consent from the patients or their legal representatives. All procedures that contributed to this work complied with the Helsinki Declaration of 1975 and its later amendments.

Clinical data We collected the following data: age, gender, comorbid

illnesses for the Charlson Comorbidity Index, sepsis origins, Simplified Acute Physiology Score II on admission, and Sequential Organ Failure Assessment (SOFA) scores on day 1 and day 3. Laboratory data, including routine chemistry and lactate, were also collected. For all enrolled patients, coagulation profiles (prothrombin time [PT], activated partial thromboplastin time, fibrinogen, and D-dimer) were collected. In terms of patient outcomes, we investigated the ICU mortality, as well as hospital and 30-day mortalities. In addition, we investigated the incidence of new thromboembolic events (stroke, myocardial infarction, ischaemic bowel disease, deep vein thrombosis and pulmonary embolism) during the hospital stay. None of the patients in this study received recombinant human activated protein C because it is not available in Korea.

DIC and SOFA scoresFor the diagnosis of DIC, we used the ISTH criteria (Table

1), which were suggested by Taylor et al in 200112. These criteria are based on four factors 1) platelet count, 2) fibrin-related products (fibrin degradation products or D-dimer), 3) PT, and 4) fibrinogen. With regard to fibrin-related markers, we used D-dimer in this study1,3. The total maximum DIC score was 8 and a score of ≥5 was defined as ‘overt DIC’. We collected blood samples to determine the DIC score from all the patients on day 1 (0 to 24 hours post-sepsis diagnosis) and day 3 (48 to 72 hours post-sepsis diagnosis). If multiple tests were performed in a single 24-hour period, the most abnormal test was used to calculate the DIC score. The PT and D-dimer were measured with a STA-REvolution® (Stago Inc., Parsippany, NJ, USA) and their normal ranges were 11.5 to 14.0 seconds and 0 to 2.7 nmol/l, respectively. SOFA scores were also calculated using the most aberrant physiologic or laboratory results on day 1 and day 315.

Data analysisPrimary outcomes were the association of day 1 and day

3 DIC scores as well as ∆DIC scores with hospital mortality. Secondary outcomes were the correlation between the DIC and SOFA-platelet (i.e. SOFA score omitting platelets) scores on day 1 and day 3, and the incidences of overt DIC by different sepsis origins (pneumonia versus urinary tract infection [UTI] versus non-pneumonia/non-UTI sepsis); we compared the incidences of overt DIC, rates of bacteraemia and hospital mortalities between these three groups.

Table 1Day 1 and day 3 coagulation profiles in survivors and non-survivors

Values Survivors (n=159) Non-survivors (n=60)

P-value

Day 1

Platelets (×109/l) 151.0 (96.0–226.0) 127.5 (64.5–219.0) 0.133

PT (s) 16.7±2.6 19.2±4.9 < 0.001

Fibrinogen (µmol/l)

16.5±6.17 15.0±5.0 0.052

D-dimer (nmol/l) 19.4 (15.7–50.8) 18.4 (11.9–52.9) 0.707

DIC score 3.8±1.1 4.3±1.3 0.010

Overt DIC 37 (23.3%) 24 (40.0%) 0.014

Day 3

Platelets (×109/l) 124.0 (82.0–195.0) 82.5 (29.3–164.3) 0.002

PT (s) 15.8±2.7 20.1±7.7 < 0.001

Fibrinogen (µmol/l)

16.5±5.8 15.3±5.6 0.121

D-dimer (nmol/l) 18.4 (11.9–35.6) 24.3 (16.2–55.6) 0.007

DIC score 3.6±1.1 4.8±1.5 < 0.001

Overt DIC 31 (19.5%) 35 (58.3%) < 0.001

PT=prothrombin time, DIC=disseminated intravascular coagulation.


59


Data are expressed as means ± standard deviations or medians (IQRs) for continuous variables and as percentages for categorical variables. Student’s t-test or one-way analysis of variance was performed for continuous data, and chi-square or Fisher’s exact tests were used for categorical data. Multivariate analyses by logistic regression were performed with covariates significant (P <0.05) in univariate analyses to investigate risk factors for hospital mortality. A backward stepwise selection method (based on the likelihood ratio) was used, with P=0.05 to enter and P=0.10 to stay in the model. Receiver operating characteristic (ROC) curves were also generated to evaluate the performance of the DIC scores for predicting hospital mortality; we compared areas under the receiver operating characteristic curves using a method suggested by Hanley et al16. All reported P-values were two-sided, and P <0.05 indicated statistical significance. All analyses were conducted using the SPSS statistical software (IBM SPSS Statistics version 21, Armonk, NJ, USA).

Results

Study populationDuring the study period, a total of 381 patients were initially

screened and among them, 219 patients were enrolled in this study (Figure 1). The mean age was 72.5 ± 12.0 years and 44.7% of the patients were female. Patients with septic shock and severe sepsis accounted for 62.6% and 21.0%, respectively and among the sepsis origins, pneumonia (n=110; 50.2%), urinary (n=51; 23.3%), and biliary infections (n=36; 16.4%) were the most common. Among comorbidities, hypertension (51.1%) and diabetes (33.8%) were the most prevalent, and the mean Charlson Comorbidity Index was 5.3 ± 2.1. One hundred and fourteen patients (52.1%) had community-acquired infections, 50 patients had healthcare-associated infections, and 55 had hospital-acquired infections. Gram-positive organisms were identified from 62 patients and Gram-negative organisms were identified from 127 patients (other organisms, n=10; no organisms identified, n=20; Table 2). Ninety-three (42.5%) patients had bacteraemia. With regard to thromboprophylaxis, 77 (35.2%) patients received unfractionated heparin, 64 (29.2%) received low molecular weight heparin and 78 (35.6%) received intermittent pneumatic compression.

DIC and SOFA scoresThe mean DIC scores on day 1 and day 3 were 4.0 ± 1.2 and

4.0 ± 1.3, respectively, and the incidences of overt DIC were 27.9% (n=61) and 30.1% (n=66), respectively. In total, 39.7% of patients (n=87) developed overt DIC during the first 72 hours. The mean SOFA scores were 6.9 ± 3.4 for day 1 and 6.6 ± 4.1 for day 3. There was a modest correlation between DIC and SOFA-platelet scores on day 3 (Figure 2).

Pneumonia sepsis compared to other sepsis originsBoth day 1 and day 3 SOFA scores were highest in patients

with pneumonia compared to UTI and non-pneumonia/non-UTI groups (day 1: 7.5 ± 3.5, 6.4 ± 3.0, 6.0 ± 3.2, respectively, P=0.012; day 3: 7.6 ± 3.9, 5.3 ± 4.2, 5.7 ± 4.0, respectively, P <0.001 [Table 3]). However, the day 1 DIC score was lowest in pneumonia patients compared to the other two groups (3.8 ± 1.1, 4.2 ± 1.4, 4.2 ± 1.2, respectively, P=0.044), but there was no difference on day 3 among the three groups. The rates of overt DIC, bacteraemia, and hospital mortality are shown in Figure 3. Pneumonia was associated with a lower incidence of overt DIC on day 1, but with a higher hospital mortality rate compared to the other groups.

381 patients screened

Cancer, 63Refusal to consent, 35Do-not-resuscitate state, 21Liver cirrhosis, 9Immunocompromised, 16CPR at admission, 7Transfer to other hospitals, 5Tuberculosis, 2

219 patients screened

Death within 72 h, 4

Figure 1: Flow diagram of enrolled patients. CPR=cardiopulmonary resuscitation.

Table 2Comparisons of clinical and laboratory variables between survivors and non-

survivors

Values Survivors (n=159) Non-survivors (n=60)

P-value

Age, yrs 72.2±12.5 73.4±10.5 0.498

Gender, female 75 (47.2%) 23 (38.3%) 0.241

CCI 5.1±2.1 5.9±1.9 0.013

CAI/HCAI/HAI 88/36/35 26/14/20 0.179

Neutropenia 6 (3.8%) 5 (8.3%) 0.178

Lactate (mmol/l) 3.4 (2.3–5.4) 3.7 (2.2–7.2) 0.457†

C-reactive protein (mg/l)

159.0 (103.5–195.9)

149.1 (107.5–221.4)

0.729

Pneumonia sepsis 64 (40.3%) 46 (76.7) < 0.001

Bacteraemia 74 (46.5%) 19 (31.7%) 0.047

Admission SAPS II 44.1±13.2 54.9±11.9 < 0.001

Predicted mortality‡ 35.0±22.5 55.5±22.0 < 0.001

SOFA score on day 1 6.1±3.1 8.9±3.2 < 0.001

SOFA score on day 3 5.1±3.3 10.4±3.7 < 0.001

∆SOFA score* 0.0 (-3.0–1.0) 1.5 (-1.0–4.0) < 0.001†

∆DIC score* -0.2±1.0 0.5±1.1 < 0.001

Mechanical ventilation

69 (43.4%) 58 (96.7%) < 0.001

Use of vasopressors 104 (65.4%) 57 (95.0%) < 0.001

CRRT 10 (6.3%) 20 (33.3%) < 0.001

*Change between day 1 and day 3 scores. †Mann–Whitney U test. ‡by admission SAPS II. CCI=Charlson Comorbidity Index, CAI=community-acquired infection, HCAI=healthcare-associated infection, HAI=hospital-acquired infec-tion, SAPS=Simplified Acute Physiology Score, SOFA=Sequential Organ Failure Assessment, DIC=disseminated intravascular coagulation, CRRT=continuous renal replacement therapy.

J. Y. Park et al

60


DIC scores and thromboembolic eventsIn total, 11 thromboembolic events (stroke, 4; myocardial

infarction, 6; ischaemic bowel disease, 0; deep vein thrombosis, 0; and pulmonary embolism, 1) occurred during the hospital stay, but with no significant difference in the incidence rate between patients with overt DIC and those without (data not shown). However, patients with thromboembolic events had a higher fibrinogen concentration (18.8 ± 5.9 µmol/l versus 15.9 ± 5.6, P=0.030) and platelet count (212×109 [122 to 294] versus 139 [91 to 217], P=0.067) on day 1, compared to those without events.

DIC scores and hospital mortalityThe ICU and hospital mortality rates were 26.0% and 27.4%,

respectively, and the 30-day mortality rate was 26.9%. Among clinical and laboratory variables, PT (day 1 and 3), platelet (day 3), D-dimer (day 3), Charlson Comorbidity Index, pneumonia sepsis, bacteraemia, admission Simplified Acute Physiology Score II, DIC scores (day 1, 3, and ∆DIC), SOFA scores (day 1, 3, and ∆SOFA), mechanical ventilation, use of vasopressors, and continuous renal replacement therapy were significantly associated with hospital mortality according to the results of the univariate analyses (Tables 1 and 2). Using these significant variables, we conducted multivariate logistic analyses. In the final model, pneumonia, mechanical ventilation, use of vasopressors, continuous renal replacement therapy, SOFA score (day 3), and ∆DIC score (odds ratio, 1.862; 95% confidence interval, 1.061 to 3.266; P=0.030) were significantly associated with increased hospital mortality (Table 3). When DIC scores were analysed as dichotomous variables (overt DIC), there was a suggestion that occurrence of overt DIC on day 3 was associated with hospital mortality, but this did not reach statistical significance (odds ratio, 2.843; 95% confidence interval, 0.844 to 9.578; P=0.092).

2 24 6 8 3 4 5 6 7

Day 1 DIC score Day 3 DIC score

Day

1 S

OFA

-pla

tele

t

0

2.5

5

7.5

10

12.5

-5

0

5

10

15

20

Day

3 S

OFA

-pla

tele

t

r=0.119P=0.078

r=0.416P <0.001

a. b.

Figure 2: Relationship between the Sequential Organ Failure Assessment without platelet (SOFA-platelet) and disseminated intravascular coagulation (DIC) scores at day 1 (a.) and day 3 (b.).

Table 3Multivariate analyses for the predictors of hospital mortality*

Beta SE P-value OR 95% CI

PT on day 3 0.063 0.056 0.266 1.065 0.953–1.189

Pneumonia sepsis

1.288 0.575 0.025 3.627 1.176–11.184

Mechanical ventilation

2.039 0.844 0.016 7.683 1.468–40.213

Use of vasopressors

1.613 0.817 0.048 5.017 1.012–24.880

CRRT 2.005 0.704 0.004 7.427 1.868–29.526

SAPS II -0.033 0.022 0.131 0.967 0.926–1.010

SOFA score on day 3

0.352 0.121 0.004 1.421 1.120–1.803

∆SOFA score -0.067 0.096 0.483 0.935 0.775–1.128

DIC score on day 3

-0.203 0.269 0.451 0.816 0.482–1.383

∆DIC score 0.622 0.287 0.030 1.862 1.061–3.266

Overt DIC on day 1†

-1.322 0.676 0.050 0.267 0.071–1.002

Overt DIC on day 3†

1.045 0.620 0.092 2.843 0.844–9.578

*Hosmer–Lemeshow test, chi-square=5.396 and P=0.715 (accuracy for discrimination, 86.8%). †Hosmer–Lemeshow test, chi-square=6.910 and P=0.546, (accuracy for discrimination 84.0%). OR=odds ratio, CI=confidence interval, PT=prothrombin time, CRRT=continuous renal replacement therapy, SAPS II=Simplified Acute Physiology Score II, SOFA=Sequential Organ Failure Assessment, DIC=disseminated intravascular coagulation, SE=standard error.


61


Comparison of day 1 and day 3 DIC scores for predicting hospital mortality

Receiver operating characteristic curves for the performances of day 1 and day 3 DIC scores are shown in Figure 4. When the area under the curve values were compared between day 1 and day 3 DIC scores, they were significantly different for all patients and patients with pneumonia sepsis or non-pneumonia sepsis. In particular, the area under the curve value for the day 3 DIC score was significantly higher in patients with non-pneumonia sepsis than in those with pneumonia origins. However, as shown in Table 4, neither DIC scores nor any combination of scores outperformed the day 3 SOFA score for all patients.

DiscussionThe current study has revealed several interesting findings.

Firstly, the ∆DIC score was significantly associated with hospital mortality in patients with sepsis who survived 72 hours. Secondly, patients with pneumonia sepsis had a lower incidence of DIC on day 1, but had a higher SOFA score and mortality rate compared to patients with other sepsis origins. Thirdly, the day 3 DIC score more accurately predicted hospital mortality than the day 1 DIC score and may be a more useful prognostic indicator for patients with non-pneumonia sepsis than for those with pneumonia sepsis.

DIC, an extreme form of systemic activation of coagulation, is characterised by widespread microvascular thrombosis, which can compromise the blood supply to various organs (tissue hypoperfusion)17. Previously, Dhainaut et al showed that worsening coagulopathy during the first day of severe sepsis was associated with new organ failure and 28-day death9. Ferreira et al also demonstrated that during the early period of ICU admission, an increased number of organ failures was associated with poor hospital outcomes18. Therefore, although the treatment for DIC is still limited to supportive care, the diagnosis and management of DIC during the early stages of sepsis seem to be critical.

In the present study, about 40% of patients developed overt DIC during the first 72 hours of sepsis, which was similar to previous reports. However, the main finding of our study was that the ∆DIC score was significantly associated with increased hospital mortality. Although the continuous variable (day 3 DIC score) did not show a significant association in the multivariate model, which could have been partly because of the small number of patients, the occurrence of overt DIC on day 3 had a marginal tendency to be associated with hospital mortality. From these results, we suggest that the evolution of the DIC profile during the early period of sepsis might reflect the host response to infection, and appropriate initial treatments are very important. In a study by Dhainaut et al, the authors used ‘composite coagulopathy score’, which included changes in antithrombin, PT, and D-dimer levels9. Kinasewitz et al also used a ‘simple evolving DIC score’, which included changes in platelet count and PT levels10. Both studies showed that worsening scores reflected poor patient outcomes, and had several aspects in common with our study.

Additionally, we determined that the correlation between the DIC and SOFA-platelet scores tended to be stronger at day 3 than at day 1, and from our ROC curves, the day 3 DIC score predicted hospital mortality better than the day 1 score. These results indicate that the day 3 DIC score may be more useful than the day 1 score with respect to disease severity and prognosis. Considering the importance of implementing the appropriate treatment and the host response early during the treatment period, these results

Figure 3: Comparison of the incidence of overt disseminated intravascular coagulation (DIC; 21.8%, 35.3%, and 32.8% on day 1; 29.1%, 35.3%, and 27.6% on day 3), bacteraemia (19.1%, 62.7%, and 69.0%), and hospital mortality (41.8%, 9.8%, and 15.5%) among patients with pneumonia, UTI (urinary tract infection), and non-pneumonia/non-UTI sepsis, respectively.

Table 4AUC values of DIC and severity scores for predicting hospital mortality

Scores AUC values 95% CI

Day 1 DIC score 0.617 0.532–0.703

Day 3 DIC score 0.733 0.654–0.812

∆DIC score* 0.674 0.594–0.755

SAPS II 0.751 0.680–0.822

SOFA score on day 1 0.737 0.665–0.809

SOFA score on day 3 0.852 0.801–0.903

∆SOFA score* 0.684 0.602–0.766

SOFA-Platelet and DIC on day 1† 0.732 0.661–0.803

SOFA-Platelet and DIC on day 3† 0.869 0.820–0.917

Day 1 SOFA and ∆DIC† 0.779 0.710–0.848

SAPS II and day 1 DIC score† 0.752 0.681–0.823

SAPS II and ∆DIC score† 0.759 0.689–0.829

*Change between day 1 and day 3 scores. †Combination of the two scores. AUC=area under the receiver operating curve, DIC=disseminated intravascu-lar coagulation, SAPS=Simplified Acute Physiology Score, SOFA=Sequential Organ Failure Assessment, SOFA-Platelet=SOFA without platelet score.

J. Y. Park et al

62


seem to be relevant. Traditionally, the lung has been one of the most common

origins of sepsis, and pneumonia sepsis has been associated with high mortality. Therefore, several authors have studied the treatment effects of coagulation inhibitors, such as antithrombin (AT) and recombinant human soluble thrombomodulin, in patients with sepsis-associated DIC of pneumonia origin19–21. Tagami et al showed in their retrospective, large, nationwide database study that AT treatment was associated with a reduction in the 28-day death rate in patients with pneumonia sepsis and DIC20. In their study, the 28-day mortality rate was 44.3%, which is similar to our result (41.8%). In our study, although pneumonia sepsis was associated with a higher disease severity (on day 1 and day 3) and higher hospital mortality rate, it was associated with a lower DIC score and a lower prevalence of overt DIC on day 1, compared to non-pneumonia sepsis patients. Although coagulation inhibitors (e.g., AT III, protein C, and protein S) were not tested in our study, it is likely that sepsis-associated coagulopathy may be subtle in pneumonia sepsis patients, or that the ISTH criteria for DIC on admission day might not accurately reflect the disease severity in cases of pneumonia sepsis. However, previously, many experts believed that the pathophysiology and mortality rate of DIC could vary depending on different sites of infection22,23. Some studies also showed a significant heterogeneity among endothelial cell phenotypes throughout the vascular tree, and organ-specific variations of haemostasis among different organs24,25.

Another finding in this study was that the DIC score was more accurate at predicting hospital deaths in non-pneumonia sepsis patients than in cases of pneumonia sepsis. In particular, the 95% confidence intervals of the day 3 DIC scores in the ROC curves did not overlap between pneumonia and non-pneumonia sepsis. Therefore, it is possible that the

ISTH criteria for DIC (day 3 DIC score) can more accurately discriminate between survivors and non-survivors in patients who acquired sepsis from non-pneumonia origins. The inferior performance of the DIC score in pneumonia sepsis could be, in part, explained by the low incidence of overt DIC, despite having a higher mortality. However, we observed that the DIC score, whether alone or in combination with other scores, was not superior to existing scoring systems.

Recent data suggest that no thromboprophylaxis within 24 hours of ICU admission is associated with an increased risk of death26. Although our study was not blinded for treatments or the results of the DIC scores, the incidence of thromboembolic events was independent of the occurrence of overt DIC. However, patients with thromboembolic events had higher fibrinogen levels on day 1, compared to those without such events. This result is consistent with the results of a study by Ho et al, showing that platelet and fibrinogen concentrations are associated with increased thrombotic and bleeding risks in vitro27. Importantly, there is also growing evidence that coagulation activation may be beneficial for bacterial clearance, as opposed to the traditional concept of DIC in sepsis (tissue damage). Some experts insist that coagulation is an essential part of the innate immune system and that appropriate haemostasis contributes to pathogen clearance by forming a physical barrier5,6. They suggest that there may be a period of time when coagulation activation changes from a beneficial to a harmful process during sepsis. Therefore, physicians should be aware of this ongoing paradigm shift to better understand the pathophysiology of DIC in sepsis. This insight may help researchers to obtain promising results in future clinical trials.

There are several limitations to this study. Firstly, our study was small and there may have been an unidentified bias. Therefore, a well-designed, large-scale study is needed to elucidate the evolution of DIC among different sepsis

1.0

0.8

0.6

0.4

0.2

0.0

Sens

itivi

ty

1.0

0.8

0.6

0.4

0.2

0.0Se

nsiti

vity

1.0

0.8

0.6

0.4

0.2

0.0

Sens

itivi

ty

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.01 – Speci�city 1 – Speci�city 1 – Speci�city

P=0.001

0.733 (0.654–0.812) vs. 0.617 (0.532–0.703)

P=0.037

0.707 (0.607–0.806) vs. 0.609 (0.499–0.718)

P=0.008

0.913 (0.846–0.979) vs. 0.785 (0.677–0.893)

a. b. c.

Figure 4: Receiver operating characteristic (ROC) curves for DIC scores for predicting hospital mortality; (a.) ROC curves for all enrolled patients, (b.) ROC curves for patients with pneumonia sepsis, (c.) ROC curves for patients with non-pneumonia sepsis. Dotted lines represent the day 1 DIC score and solid lines represent the day 3 DIC score. DIC=disseminated intravascular coagulation.


63


origins. Secondly, as mentioned above, we did not test other coagulation inhibitors, such as AT III, protein C and protein S. These tests, if used, would have strengthened the robustness of our study. Finally, unlike other studies that included only patients with severe sepsis or septic shock, we included patients with sepsis, which might explain why we could not demonstrate the significance of DIC in some of the outcomes in this study. However, the strength of our study is that we collected data from sepsis patients prospectively and demonstrated several meaningful findings regarding the DIC score, after controlling for important confounders (pre-existing severity scores) in the multivariate analysis.

ConclusionsIn conclusion, the ∆DIC score was significantly associated

with hospital mortality in patients with sepsis. Specifically, patients with pneumonia sepsis had a lower incidence of DIC on day 1 despite their higher disease severity and mortality rates, compared to those with other sepsis origins. In addition, the day 3 DIC score predicted hospital deaths more accurately than the day 1 score. The day 3 DIC score might be a more useful indicator for patients with non-pneumonia sepsis than for those with pneumonia. However, the DIC score did not provide additional discriminative power over existing scoring systems.

AcknowledgementThe authors would like to thank Young-Su Joo (Department

of Occupational and Environmental Medicine) and Sun-Ho Lee (Department of Internal Medicine) for their contribution to the statistical analyses and data collection.

Funding This study was supported by a grant from Hallym University

Medical Center Research Fund (01-2010-05).

Editor’s noteSupplementary material accompanying this paper can be

accessed in the online version of the article.

References1. Bakhtiari K, Meijers JC, de Jonge E, Levi M. Prospective valida-

tion of the international society of thrombosis and haemostasis scoring system for disseminated intravascular coagulation. Crit Care Med 2004; 32:2416-2421.

2. Dhainaut JF, Yan SB, Joyce DE, Pettila V, Basson B, Brandt JT et al. Treatment effects of drotrecogin alfa (activated) in patients with severe sepsis with or without overt disseminated intravascular coagulation. J Thromb Haemost 2004; 2:1924-1933.

3. Levi M. Disseminated intravascular coagulation. Crit Care Med 2007; 35:2191-2195.

4. Iba T, Nagaoka I, Boulat M. The anticoagulant therapy for sepsis-associated disseminated intravascular coagulation. Thromb Res 2013; 131:383-389.

5. van der Poll T, Herwald H. The coagulation system and its function in early immune defense. Thromb Haemost 2014; 112:640-648.

6. Fiusa MM, Carvalho-Filho MA, Annichino-Bizzacchi JM, De Paula EV. Causes and consequences of coagulation activation in sepsis: an evolutionary medicine perspective. BMC Med 2015; 13:105-113.

7. Kinasewitz GT, Yan SB, Basson B, Comp P, Russell JA, Cariou A et al. Universal changes in biomarkers of coagulation and inflammation occur in patients with severe sepsis, regardless of causative micro-organism [ISRCTN74215569]. Crit Care 2004; 8:R82-R90.

8. Nijsten MW, ten Duis HJ, Zijlstra JG, Porte RJ, Zwaveling JH, Paling JC et al. Blunted rise in platelet count in critically ill patients is associated with worse outcome. Crit Care Med 2000; 28:3843-3846.

9. Dhainaut JF, Shorr AF, Macias WL, Kollef MJ, Levi M, Reinhart K et al. Dynamic evolution of coagulopathy in the first day of severe sepsis: relationship with mortality and organ failure. Crit Care Med 2005; 33:341-348.

10. Kinasewitz GT, Zein JG, Lee GL, Nazir SA, Taylor FB, Jr. Prognostic value of a simple evolving disseminated intravascular coagula-tion score in patients with severe sepsis. Crit Care Med 2005; 33:2214-2221.

11. MacKenzie IM. The haemodynamics of human septic shock. Anaesthesia 2001; 56:130-144.

12. Taylor FB, Jr, Toh CH, Hoots WK, Wada H, Levi M. Towards defi-nition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation. Thromb Haemost 2001; 86:1327-1330.

13. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, Jaeschke R et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2008. Crit Care Med 2008; 36:296-327.

14. Friedman ND, Kaye KS, Stout JE, McGarry SA, Trivette SL, Briggs JP et al. Health care--associated bloodstream infections in adults: A reason to change the accepted definition of communi-ty-acquired infections. Ann Intern Med 2002; 137:791-797.

15. Vincent JL, de Mendonca A, Cantraine F, Moreno R, Takala J, Suter PM et al. Use of the SFOA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on "sepsis-related problems" of the European Society of Intensive Care Medicine. Crit Care Med 1998; 26:1793-1800.

16. Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983; 148:839-843.

17. Levi M, Ten Cate H. Disseminated intravascular coagulation. N Engl J Med 1999; 341:586-592.

18. Ferreira FL, Bota DP, Bross A, Melot C, Vincent JL. Serial evaluation of the SOFA score to predict outcome in critically ill patients. JAMA 2001; 286:1754-1758.

19. Gando S, Saitoh D, Ishikura H, Ueyama M, Otomo Y, Oda S et al. A randomized, controlled, multicenter trial of the effects of antithrombin on disseminated intravascular coagulation in patients with sepsis. Crit Care 2013; 17:R297-R306.

J. Y. Park et al

64


20. Tagami T, Matsui H, Horiguchi H, Fushimi K, Yasunaga H. Antithrombin and mortality in severe pneumonia patients with sepsis-associated disseminated intravascular coagulation: an observational nationwide study. J Thromb Haemost 2014; 12:1470-1479.

21. Tagami T, Matsui H, Horiguchi H, Fushimi K, Yasunaga H. Recombinant human soluble thrombomodulin and mortality in severe pneumonia patients with sepsis-associated disseminated intravascular coagulation: an observational nationwide study. J Thromb Haemost 2015; 13:31-40.

22. Gando S, Saitoh D, Ogura H, Mayumi T, Koseki K, Ikeda T et al. Natural history of disseminated intravascular coagulation diag-nosed based on the newly established diagnostic criteria for critically ill patients: results of a multicenter, prospective survey. Crit Care Med 2008; 36:145-150.

23. Wada H, Asakura H, Okamoto K, Iba T, Uchiyama T, Kawasugi K et al. Expert consensus for the treatment of disseminated intravas-cular coagulation in Japan. Thromb Res 2010; 125:6-11.

24. Rosenberg RD, Aird WC. Vascular-bed--specific hemostasis and hypercoagulable states. N Engl J Med 1999; 340:1555-1564.

25. Monahan-Earley R, Dvorak AM, Aird WC. Evolutionary origins of the blood vascular system and endothelium. J Thromb Haemost 2013; 11 Suppl 1:46-66.

26. Ho KM, Chavan S, Pilcher D. Omission of early thromboprophy-laxis and mortality in critically ill patients: A multicenter registry study. Chest 2011; 140:1436-1446.

27. Ho KM, Duff OC. Predictors of an increased in vitro thrombotic and bleeding tendency in critically ill trauma and non-trauma patients. Anaesth Intensive Care 2015; 43:317-322.

Copyright of Anaesthesia & Intensive Care is the property of Australian Society ofAnaesthetists and its content may not be copied or emailed to multiple sites or posted to alistserv without the copyright holder's express written permission. However, users may print,download, or email articles for individual use.

Health & Medicine

Day 3 versus day 1 disseminated intravascular coagulation score among sepsis patients