Advertisement

The ISTH DIC score predicts outcome in non-septic patients admitted to a cardiovascular intensive care unit

Open AccessPublished:July 01, 2020DOI:https://doi.org/10.1016/j.ejim.2020.06.017

      Abstract

      Background

      The International Society of Thrombosis and Haemostasis (ISTH) disseminated intravascular coagulation (DIC) score is widely used to predict mortality in critically ill - typically septic - patients. The objective of this study was to investigate whether the ISTH DIC-2001 and DIC-2018 score can be used to predict the 30-day mortality in non-septic patients in an intensive care unit (ICU).

      Methods

      In this single-center, prospective observational study we included all patients ≥18 years of age who were admitted to a medical ICU with a focus on cardiovascular diseases between August 2012 and 2013. The DIC-2001 and DIC-2018 scores were calculated on admission (DIC-2001-0h and DIC-2018-0h) and 72 hours thereafter (DIC-2001-72h and DIC-2018-72h) and were classified as overt when ≥ 5 for DIC-2001 and ≥ 4 for DIC-2018.

      Results

      A total of 233 patients were included in this study. Excluding septic patients and patients after routine surgery/procedures, we calculated the DIC score for 167 patients (32.4% female; median age 64.9 years). Overt DIC-2001-0h, DIC-2018-0h and overt DIC-2001-72h scores were associated with a significantly higher 30-day mortality rate (52.9% vs. 25.0%, 46.2% vs 21.2%, and 57.1% vs. 23.7%; p < 0.04). The DIC-2001 scores and the DIC-2018-0h score significantly predicted the 30-day mortality.

      Conclusion

      This study suggests that the DIC score may be applied to non-septic ICU populations, and indicates that it is a useful tool for mortality prediction, regardless of the underlying disease.

      Keywords

      1. Introduction

      The International Society of Thrombosis and Haemostasis (ISTH) disseminated intravascular coagulation (DIC) score is used to predict mortality in various critical conditions, including severe sepsis and obstetric emergencies [
      • Yamakawa K.
      • et al.
      Recombinant human soluble thrombomodulin in severe sepsis: a systematic review and meta-analysis.
      ,
      • Gando S.
      • Levi M.
      • Toh C.H.
      Disseminated intravascular coagulation.
      ]. It consists of prothrombin time, platelet count, fibrinogen and D-dimer levels, assessed on an 8-point scale. At the time this study was conducted, a score ≥ 5 indicated an overt DIC score [
      • Toh C.H.
      • Hoots W.K.
      S.S.C.o.D.I.C.o.t. ISTH
      The scoring system of the scientific and standardisation committee on disseminated intravascular coagulation of the international society on thrombosis and haemostasis: a 5-year overview.
      ]. In 2018, the cut-off values for both D-dimer levels and an overt DIC score were changed in order to allow for higher sensitivity and specificity [
      • Suzuki K.
      • et al.
      A re-evaluation of the D-dimer cut-off value for making a diagnosis according to the ISTH overt-DIC diagnostic criteria: communication from the SSC of the ISTH.
      ].
      The score was initially only used if there was the diagnosis of an underlying condition associated with DIC [
      • Taylor Jr., F.B.
      • et al.
      Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation.
      ]. Studies regarding the ISTH DIC score focus mainly on patients with sepsis [
      • Yamakawa K.
      • et al.
      Recombinant human soluble thrombomodulin in severe sepsis: a systematic review and meta-analysis.
      ,
      • Gando S.
      • Levi M.
      • Toh C.H.
      Disseminated intravascular coagulation.
      ]. However, there is emerging evidence that the ISTH DIC score predicts the mortality in populations other than those previously assumed, for example in patients who have suffered cardiac arrest [
      • Buchtele N.
      • et al.
      Added value of the DIC score and of D-dimer to predict outcome after successfully resuscitated out-of-hospital cardiac arrest.
      ] or those who have had cardiopulmonary bypass surgery [
      • Demma L.J.
      • et al.
      Predicting mortality in patients with disseminated intravascular coagulation after cardiopulmonary bypass surgery by utilizing two scoring systems.
      ]. We hypothesized that the DIC score may also be predictive in non-septic intensive care unit (ICU) patients admitted to a medical ICU.
      In this study we therefore investigated whether the ISTH DIC score can also be applied to predict 30-day mortality in unselected, non-septic patients at a medical, cardiovascular ICU.

      2. Methods

      2.1 Data collection

      This study was approved by the independent Ethics Committee of the Medical University of Vienna and complies with the Declaration of Helsinki. All patients ≥ 18 years admitted to the medical ICU of the Department of Internal Medicine II at the General Hospital of Vienna between August 2012 and August 2013 were included consecutively. Two biomarker papers investigating this population have recently been published, describing the data collection in full detail [
      • Krychtiuk K.A.
      • et al.
      Monocyte subset distribution is associated with mortality in critically ill patients.
      ,
      • Krychtiuk K.A.
      • et al.
      Mitochondrial DNA and Toll-Like Receptor-9 Are Associated With Mortality in Critically Ill Patients.
      ]. This ICU treats the full spectrum of critically ill medical patients and has a particular focus on cardiovascular diseases; however, it also admits patients following cardiothoracic surgery. Patients admitted after routine surgery or routine procedures (such as transcatheter aortic valve implantation and percutaneous edge-to-edge mitral valve repair) as well as patients diagnosed with sepsis were excluded.

      2.2 Blood sampling

      Blood was drawn via an arterial or central venous line within 24 hours of admission, as well as 72 hours after this first blood draw. Analytical methods have been described previously [
      • Schwameis M.
      • et al.
      Asphyxia by Drowning Induces Massive Bleeding Due To Hyperfibrinolytic Disseminated Intravascular Coagulation.
      ].

      2.3 DIC Score

      The ISTH DIC 2001 score was calculated as follows [
      • Toh C.H.
      • Hoots W.K.
      S.S.C.o.D.I.C.o.t. ISTH
      The scoring system of the scientific and standardisation committee on disseminated intravascular coagulation of the international society on thrombosis and haemostasis: a 5-year overview.
      ]: A platelet count >100 G/L accounts for 0 points, a score between 50-100 G/L for 1 point and a platelet count <50 G/L for 2 points. Prothrombin time, given in %, results in 0 points when >70%, 1 point when 40-70% and 2 points when <40%. Fibrinogen levels >100 mg/dL result in 0 points, <100 mg/dL in 1 point. D-dimer levels of <1 µg/mL account for 0 points, levels between 1 and 2 µg/mL account for 2 points and D-dimer levels >2 µg/mL account for 3 points. The ISTH DIC 2001 score was calculated on admission (DIC-2001-0h) and 72 hours thereafter (DIC-2001-72h). A sum score of ≥ 5 is defined as an overt DIC-2001 score. The ISTH DIC-2018 scores (DIC-2018-0h and DIC-2018-72h) were calculated using different cut-off values for D-dimer levels (0 point < 3.0 µg/mL; 2 points 3.0 to 7.0 µg/mL; 3 points > 7.0 µg/mL). Other cut-off values were the same. A score sum of ≥ 4 is defined as an overt DIC-2018 score [
      • Suzuki K.
      • et al.
      A re-evaluation of the D-dimer cut-off value for making a diagnosis according to the ISTH overt-DIC diagnostic criteria: communication from the SSC of the ISTH.
      ].
      The data collection was carried out prospectively. However, the calculation of the DIC score was performed retrospectively. Thus, the DIC score did not affect the treatment of patients in this study.

      2.4 Statistical analysis

      Categorical variables are summarized as counts and percentages and are compared using the χ2 or Fisher's exact test as appropriate. Continuous variables are expressed as median and interquartile range (IQR). Univariate differences between groups were assessed using the Mann-Whitney U test. Cox proportional hazard regression analysis was performed to assess the effect an overt DIC score and other factors on mortality. We selected covariables for the multivariable models based on both clinical reasoning and previous studies. Kaplan–Meier analysis (log-rank test) was applied to verify the time-dependent discriminative power of the calculated overt DIC score. No imputation for missing data was performed. Two-sided p-values of <0.05 indicated statistical significance. SPSS 18.0 (IBM Corporation, Armonk, NY, USA), R (Version 4.0.0) and GraphPad Prism 8.4.2 were used for all analyses.

      3. Results

      A total of 233 patients were admitted to the ICU between August 2012 and August 2013. The exclusion of septic patients and patients who were admitted after routine procedures (such as transcatheter aortic valve implantation and percutaneous edge-to-edge mitral valve repair) resulted in a sample size of 183 patients. Of those patients, we were able to calculate the DIC-2001-0h and DIC-2018-0h score for 167 patients and the DIC-2001-72h and DIC-2018-72h score for 125.
      Table 1 shows the baseline characteristics of the 167 patients (32.2 % female; median age 64.9 years, IQR 54.6-74.5). The overall 30-day mortality was 28.1% (n=47). Within the first 72 hours, 10 patients died (i.e. 6.0%). 73.1% of the patients were admitted for medical reasons and 26.9% following surgery. The median stay on the ICU was 6 days (IQR 3-13). Table 2 shows the underlying cause for admission and the frequency of overt DIC scores for both the DIC-2001 and DIC-2018 score. Fig. 1 shows the differences between the DIC-2001 and DIC-2018 score on admission and 72 hours thereafter. The effects of the revision of the cut-off levels for D-dimer can be seen in Fig. 1A and 1B, and the changes in the total score can be seen in Fig. 1C.
      Table 1Baseline characteristics of the study population.
      ISTH DIC-0h-2001ISTH DIC-0h-2018
      Total

      (n=167)
      Survivor

      (n=120)
      Non-survivor

      (n=47)
      < 5

      (n=151)
      ≥ 5

      (n=16)
      < 4

      (n=147)
      ≥ 4

      (n=20)
      Age64.9 (54.6-74.5)63.7 (52.2-72.3)70.7 (59.2-78.4)64.9 (55.0-74.7)64.9 (49.8-72.9)64.9 (55.0-72.7)65.5 (49.3-72.2)
      Female gender, n (%)54 (32.2)40 (33.3)14 (29.8)48 (31.8)6 (37.5)48 (32.7)6 (30.0)
      Length of stay (days)6 (3-13)5.5 (3-13.5)9.0 (3.0-13.0)6 (3-13)8 (4-13.5)6 (3-13)8 (4-13)
      Mean Blood Pressure63 (58-71)64 (59-72)62 (57-69)63 (58-71)63 (59-68)63 (59-71)63 (56-70)
      Bilirubin (mg/dL)1.0 (0.6-1.6)0.9 (0.5-1.4)1.0 (0.6-1.8)0.9 (0.5-1.3)2.5 (1.3-5.4)0.85 (0.5-1.4)1.7 (1.0-2.5)
      Creatinine (mg/dL)1.1 (0.9-1.9)1.0 (0.8-1.7)1.6 (1.1-2.8)1.1 (0.9-1.8)1.5 (0.9-2.8)1.1 (0.9-1.81)1.5 (1.0-2.8)
      C-reactive protein (mg/dL)4.4 (1.6-9.9)4.2 (1.7-10.3)4.4 (1.6-8.7)4.4 (1.5-10.0)3.9 (2.3-6.0)4.4 (1.6-9.5)4.4 (2.1-14.6)
      BNP (pg/mL)2,697 (813-8,688)2,094 (611-6,005)8,386 (2,160-29,255)2,396 (669-8,567)6,034 (1,652-14,334)2,377 (669-8,486)7,131 (2,862-23,718)
      GCS14 (3-15)15 (3-15)3 (3-14)14 (3-15)14 (3-15)15 (3-15)11 (3-14)
      Vasopressor use, n (%)108 (64.7)68 (56.7)40 (85.1)96 (63.6)12 (75)92 (62.6)16 (80)
      Mechanical ventilation, n (%)113 (67.7)76 (63.3)37 (78.7)101 (66.9)12 (75)98 (66.7)15 (75)
      Renal replacement therapy7/144 (4.9)4 (3.6)3 (9.4)7 (5.2)0 (0)7 (5.3)0 (0)
      APACHE II score21 (13-25)17 (12-23)27 (23-30)21 (13-25)25 (18-28)20 (13-25)25 (18-30)
      Platelet count (G/L)170 (115-217)175 (122-218)160 (105-203)181 (127-223)75 (37-131)177 (122-223)121 (61-190)
      Quick-Time in %71.0 (56.0-82.0)72 (58.0-84.5)66.0 (43.0-78.0)72.0 (59.0-84.0)31.0 (22.5-56.5)73.0 (59.0-84.0)44.0 (25.0-62.0)
      Fibrinogen (mg/dL)389 (281-504)403 (297-510)350 (272-446)398 (305-513)222 (170-320)392 (305-504)275 (185-532)
      D-dimer (µg/mL)1.9 (0.8-3.3)1.52 (0.7-3.1)2.6 (1.6-5.1)1.7 (0.7-3.1)3.7 (2.5-21.0)1.6 (0.7-2.9)9.0 (3.9-24.9)
      Numbers are n (%) or median (interquartile range). ISTH … International Society on Thrombosis and Haemostasis; DIC … Disseminated Intravascular Coagulation; DIC-0h-2001… DIC score on admission using cut-off scores published in 2001; DIC-0h-2018 … DIC score on admission using cut-off scores published in 2018; APACHE II … Acute Physiology and Chronic Health Evaluation Score II; BNP … brain natriuretic peptide; GCS … Glasgow coma scale
      Table 2Cause for admission.
      ISTH-2001ISTH-2018
      n (%)DIC-0h ≥ 5

      n
      DIC-72h ≥5

      n
      DIC-0h ≥ 4

      n
      DIC-72h ≥ 4

      n
      Acute heart failure18 (10.8)-1--
      Acute renal failure2 (1.2)----
      Hemorrhagic shock4 (2.4)2-2-
      Cardiac surgery (post-surgery)45 (26.9)3637
      Cardiogenic shock24 (14.4)5365
      Cardiopulmonary resuscitation49 (29.3)4161
      Chronic obstructive pulmonary disease6 (3.6)--1-
      Insult / Intracranial bleeding2 (1.2)---1
      Miscellaneous3 (1.8)----
      Pneumonia/ Acute respiratory distress syndrome8 (4.8)-111
      Pulmonary embolism2 (1.2)11-1
      Heart transplant (post-surgery)4 (2.4)1-1-
      Total16716132016
      ISTH … International Society on Thrombosis and Haemostasis; DIC … Disseminated Intravascular Coagulation; DIC-0h-2001… DIC score on admission using cut-off scores published in 2001; DIC-0h-2018 … DIC score on admission using cut-off scores published in 2018
      Fig 1
      Fig. 1Effects of the revision of the ISTH DIC score on scores of the component D-dimer and the total DIC score. ISTH International Society on Thrombosis and Haemostasis; DIC Disseminated Intravascular Coagulation
      The rate of DIC-2001 scores ≥5 was 9.6% (95%CI: 4.0-14.4%) for the DIC-2001-0h and 10.4% (95%CI: 5.6-16.0%) for the DIC-2001-72h. For the DIC-2018-0h and DIC-2018-72h the rate of a score ≥4 was 12.0% (95%CI: 7.2-17.4%) and 12.8% (95%CI: 7.2-19.2%) respectively. In 16 patients, the DIC-2001 score changed within the first three days: the score improved from an overt DIC-2001 score to a DIC-2001 score < 5 in 7 patients, whereas 9 patients developed an overt DIC-2001 score. The ISTH DIC-2018 score changed in 15 patients. The score improved in 5 patients from an overt DIC-2018 score to a DIC-2018 score < 4. Ten patients developed an overt DIC-2018 score. There was no significant difference in the DIC-0h rate between medical (10.7% DIC-2001-0h, 13.9% DIC-2018-0h) and surgical patients (6.7% DIC-2001-0h, 6.7% DIC-2018-0h, p=0.562 and p=0.284 respectively). Furthermore, there was no significant difference in the rate of an overt DIC score 72 hours after admission (7.8% DIC-2001-72h, 10.0% DIC-2018-72h for medical patients and 20.0% DIC-2001-72h, 18.4% DIC-2018-72h for surgical patients, p=0.188 and p=0.145 respectively).
      Non-survivors showed an overt DIC-2001-0h score (19.1% vs. 5.8%, p = 0.016) and overt DIC-2018-0h score (25.5% vs. 6.7%, p=0.002) more frequently than survivors. In the remaining 157 patients, the DIC-2001-72h score (20.0% vs. 7.4%, p=0.080) and DIC-2018-72h score (20.0% vs. 10.5%, p=0.211) trended towards a higher overt score rate in non-surviving patients.
      The Acute Physiology and Chronic Health Evaluation (APACHE)-II score calculated on admission was higher in non-survivors (27 vs. 17, p < 0.001). There was no significant difference in the APACHE-II score between patients with or without an overt DIC-2001 score on admission (25 vs. 21 p = 0.094). However, the APACHE-II score was significantly higher in the patients who scored an overt DIC-2018 score on admission compared to those who did not (25 vs. 20, p = 0.021). The 30-day mortality rate in surgical patients was significantly lower than in medical patients (8.5% vs. 34.2%, p < 0.001).
      Vasopressor use was higher in non-survivors on admission (85.1% vs. 56.7%, p=.001) and 72 hours thereafter (71.9% vs. 27.6%, p<.001). Furthermore, the rate of mechanical ventilation was not significantly higher in non-survivors on admission (78.7% vs. 63.3%, p= 0.056) but significantly higher 72 hours later (87.5% vs. 32.8%, p<0.001). The rate of renal replacement therapy – recorded only at admission – did not differ significantly between survivors and non-survivors (3.6% vs. 9.4%, p = 0.184). Age differed significantly between survivors and non-survivors (63.7 vs. 70.7, p=0.010). There was no sex difference between survivors and non-survivors (33.3% vs. 29.8% female, p = 0.660).
      The presence of a DIC-2001 score ≥ 5 was associated with a significantly higher 30-day mortality rate on admission (56.3% vs. 25.2%% p=0.005 Fig. 2A). The same was true for a DIC-2001-72h score ≥5 (46.2% vs 21.4%, p=0.035, Fig. 2B) and a DIC-0h-2018 score ≥4 (60.0% vs. 23.8%, p<0.001, Fig. 2C). However, there was only a trend for a DIC-2018-72h score ≥ 4 (37.5% vs. 21.8%, p=0.16, Fig. 2D).
      Fig 2
      Fig. 2A-D. Kaplan-Meier plot of the estimated mortality in the intensive care unit (ICU) at day 30 according to the presence of an overt disseminated intravascular coagulation (DIC) score calculated for all patients by the DIC-2001 and DIC-2018 score of the International Society of Thrombosis and Haemostasis (ISTH) at admission (A: DIC-2001-0h; C: DIC-2018-0h) and 72 hours thereafter (B: DIC-2001-72h; D: DIC-2018-72h).
      The results of the Cox Regression show that a DIC-2001-0h and the DIC-2001-72h score ≥ 5 significantly predicted mortality (Table 3). This was regardless of age, sex, the APACHE II score, type of admission (medical/surgical), usage of vasopressors and mechanical ventilation on admission. The hazard ratio for a DIC-2001-0h score ≥ 5 was 2.44 (95%CI: 1.16-5.13) and for a DIC-2001-72h score ≥ 5 was 2.80 (95%CI: 1.04-7.54). In addition to the DIC-2001 score ≥ 5 and the APACHE-II score, the presence of mechanical ventilation significantly predicted the 30-day mortality when calculated 72 hours after admission. For the DIC-2018-0h score the result was similar. However, a DIC-2018-72h score ≥ 4 only showed a trend towards the prediction of mortality (HR 2.32, 95%CI 0.83 - 6.52, p=0.110).
      Table 3Cox regression
      ISTH-2001 DIC scoreISTH-2018 DIC score
      Admission72h thereafterAdmission72h thereafter
      HR (95 % CI)p-ValueHR (95 % CI)p-ValueHR (95 % CI)p-ValueHR (95 % CI)p-Value
      Age1.01 (0.99 - 1.03)0.4251.02 (0.994 – 1.05)0.1211.01 (0.99 - 1.03)0.4941.02 (0.99 – 1.05)0.122
      Sex (m/f)0.83 (0.43 - 1.61)0.5811.16 (0.53 – 2.57)0.7090.90 (0.47 – 1.75)0.7661.20 (0.54-2.66)0.651
      Overt DIC score2.44 (1.16 – 5.13)0.0192.80 (1.04 – 7.54)0.0422.10 (1.05 – 4.19)0.0372.32 (0.83-6.52)0.110
      APACHE II1.12 (1.06 - 1.18)< 0.0011.07 (1.01 – 1.14)0.0311.10 (1.05 – 1.16)< 0.0011.08 (1.01 – 1.15)0.021
      Admission (Medical/Surgery)0.43 (0.15 - 1.27)0.1250.50 (0.15 – 1.54)0.2530.40 (0.14 – 1.18)0.0970.54 (0.17 – 1.75)0.307
      Vasopressor (y/n)1.94 (0.78 – 4.83)0.1551.36 (0.57 - 3.24)0.4842.00 (0.80 – 4.98)0.1371.34 (0.56 – 3.19)0.509
      Mech. Ventilation (y/n)0.62 (0.27 – 1.45)0.2704.72 (1.21 – 18.48)0.0260.66 (0.29 – 1.153)0.3324.62 (1.18 – 18.03)0.028
      ISTH … International Society on Thrombosis and Haemostasis; DIC … Disseminated Intravascular Coagulation; DIC-0h-2001… DIC score on admission using cut-off scores published in 2001; DIC-0h-2018 … DIC score on admission using cut-off scores published in 2018; APACHE II … Acute Physiology and Chronic Health Evaluation Score II
      Additionally, we calculated a variable called “AnyDIC-2001” and “AnyDIC-2018” including patients who showed an overt DIC-2001 or DIC-2018 score at any time (i.e. an over DIC score on admission and/or 72 hours thereafter). AnyDIC-2001 also significantly predicted mortality, using the Cox regression model shown in Table 3 (HR: 2.57, 95% CI 1.15-5.74, p =0.022). We demonstrated a trend for AnyDIC-2018’s prediction of mortality(HR: 2.19, 95%CI 0.94-5.13, p=0.071).

      4. Discussion

      This study assessed the predictive value of the ISTH DIC score regarding mortality in non-septic patients admitted to a cardiovascular ICU. For this, we calculated the score used at the time this study was conducted (DIC-2001) and the revised score (DIC-2018). The DIC-2001 score predicted the 30-day mortality rate both at admission and 72 hours thereafter. This was independent of age, sex, APACHE-II score, type of admission, usage of vasopressors or mechanical ventilation. The same was true for the DIC-2018 score at admission. The DIC-2018 score calculated 72 hours after admission showed a trend towards predicting mortality. This demonstrates the usefulness of the score in a population without any underlying disorders associated with DIC.
      Repeated calculation of the DIC score may be important for several reasons: (1) the DIC-2001 score changed over time in 16 of our patients (i.e. 80% of our 20 patients with a DIC-2001 score ≥5 at any time), the DIC-2018 score changed in 15 of our patients (i.e. 71.4% of the 21 patients with a DIC-2018 score ≥ 4 at any time), (2) the DIC-2001-72h score predicted mortality, the DIC-2018-72h revealed a trend and (3) the hazard ratio for the 30-day mortality even increased between DIC-2001-0h and DIC-2001-72h. The non-significant difference between non-survivors and survivors in the prevalence of an overt DIC score for both DIC-2001-72h and DIC-2018-72h may be explained by the limited sample size and/or the patients who died within the first 72 hours. This may also be the reason for the non-significant 30-day mortality prediction of the DIC-2018-72h score.
      The prevalence of an overt DIC score in our study corresponds well with the previously reported average incidence rate of 10.6% when calculated in various underlying disorders [
      • Gando S.
      • Levi M.
      • Toh C.H.
      Disseminated intravascular coagulation.
      ] and 18.1% in severe sepsis [
      • Gando S.
      • et al.
      A multicenter, prospective validation study of the Japanese Association for Acute Medicine disseminated intravascular coagulation scoring system in patients with severe sepsis.
      ]. There was a numerical difference in the prevalence of overt DIC between medical and surgical patients, which missed statistical significance, possibly due to the small sample size.
      The higher prevalence of an overt DIC score in non-surgical patients seems unexpected. A rise in D-dimer level is a physiological response to surgery [
      • Dindo D.
      • et al.
      Kinetics of D-dimer after general surgery.
      ], which consequently could increase the rate of an overt DIC score between the DIC score calculated on admission and the DIC score calculated thereafter. Of note, a D-dimer level > 2 µg/mL accounts for two points in the DIC-2001 score. In our small sample of 46 surgical patients, 7 of 8 patients with a DIC-2001 score ≥ 5 already had a D-dimer level >2 µg/mL on ICU admission and all of them demonstrated elevated D-dimer levels after 72 hours. The decrease in platelet count could be the result of the usage of a heart-lung-machine (HLM) during surgery [
      • Sniecinski R.M.
      • Chandler W.L.
      Activation of the hemostatic system during cardiopulmonary bypass.
      ], or the physiological response to surgery itself. The increase in the DIC-2001 score in surgical patients was mainly due to a decrease in the prothrombin time. In the DIC-2018 score, only one surgical patient received three points for the D-dimer level on admission, and only three patients were assigned three points after 72 hours. This reflects the effect of the change in the cut-off values.
      The mortality rate in surgical patients was lower than in medical patients. This is somewhat expected because the majority of the surgical patients underwent elective cardiac or thoracic surgery. The mortality rate in patients with a DIC-2001-0h score ≥ 5 or DIC-2001-72h score ≥ 5 was ~3-fold higher compared to the patients without overt DIC scores. The same was true for a DIC-2018 score ≥ 4 compared with a DIC-2018 score < 4. The adjusted hazard ratios for mortality indicated a consistent, more than two-fold, risk of dying within 30 days if an overt DIC-2001 score was present on admission or 72 hours thereafter. The same was true for a DIC-2018-0h score ≥ 4. The clinical relevance of this result is highlighted even when compared to the hazard ratio of the APACHE-II score on admission. Surprisingly, there was no significant difference in the APACHE-II score between the patients with or without a DIC-2001-0h score ≥ 5, but for patients with a DIC-2018-0h score ≥ 4. The APACHE-II score was developed to quantify the severity of an illness [
      • Frost S.A.
      • et al.
      Severity of illness and risk of readmission to intensive care: a meta-analysis.
      ], which can differ between patients with the same diagnosis. In our study overt DIC scores on admission were able to predict mortality independently of the APACHE-II score. Hence, the DIC score may provide additional, clinically relevant information.
      The observed increased mortality is in agreement with studies investigating the ISTH DIC score after cardiac arrest [
      • Buchtele N.
      • et al.
      Added value of the DIC score and of D-dimer to predict outcome after successfully resuscitated out-of-hospital cardiac arrest.
      ,
      • Demma L.J.
      • et al.
      Predicting mortality in patients with disseminated intravascular coagulation after cardiopulmonary bypass surgery by utilizing two scoring systems.
      ] or cardiopulmonary bypass surgery [
      • Demma L.J.
      • et al.
      Predicting mortality in patients with disseminated intravascular coagulation after cardiopulmonary bypass surgery by utilizing two scoring systems.
      ], in the general population in an emergency department [
      • Schwameis M.
      • et al.
      Prognosis of overt disseminated intravascular coagulation in patients admitted to a medical emergency department.
      ] or in sepsis [
      • Gando S.
      • et al.
      Role of disseminated intravascular coagulation in severe sepsis.
      ]. Using sensitivity analyses (such as “AnyDIC2001” and “AnyDIC2018”) we showed that the DIC score is a robust tool for mortality prediction. This further supports the potential value of repeatedly calculating a patients’ DIC score, which may merit further investigation. One possible explanation for the high mortality rate in patients with overt an DIC score could be the underlying diseases. Most of the overt DIC score events were associated with high mortality medical conditions (e.g. cardiogenic shock, cardiopulmonary resuscitation).
      The use of the DIC score in a population without an underlying disease associated with DIC may seem counterintuitive, especially considering that this a requirement for the usage of the score [
      • Taylor Jr., F.B.
      • et al.
      Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation.
      ]. It is important to note, however, that the goal of this study was to investigate whether the DIC score is able to predict mortality in non-septic patients, not to make a DIC diagnosis. Furthermore, a score indicating an overt DIC does not mean the patient suffers from DIC. The DIC score incorporates widely available global coagulation markers. However, changes in these markers not only represent changes in coagulation but may also be influenced by the impairment of other organ systems. For example, low fibrinogen levels or altered prothrombin time may reflect impaired liver synthesis. Similarly, increased D-dimer levels may reflect the body's fibrinolytic response to hypoxemia [
      • Buchtele N.
      • et al.
      Added value of the DIC score and of D-dimer to predict outcome after successfully resuscitated out-of-hospital cardiac arrest.
      ,
      • Schwameis M.
      • et al.
      Asphyxia by Drowning Induces Massive Bleeding Due To Hyperfibrinolytic Disseminated Intravascular Coagulation.
      ]. This shows that the DIC score is also affected by severe disorders other than the disorders typically associated with DIC. It is crucial to point out that the DIC score was associated with a higher mortality, but that there was no evidence of true coagulation activation leading to DIC.
      This study is limited as it was a single-center study of one-year duration resulting in a limited sample size and low number of events. Future studies should include multiple centers and therefore expand the non-septic population investigated. Although the patient population was heterogeneous, this study provides a real-life depiction of a medical ICU with a cardiovascular focus. Unfortunately, data on the exact cause of death is missing. The majority of critical ill patients die from multiple organ failure; the precise cause of death can only be determined by autopsy. However, autopsies are only performed in very rare cases at our institution. Future studies should try to include this parameter at best through autopsy results or, alternatively, by an independent expert committee. Future studies should also investigate whether other causes of variability can be identified in patients (e.g. platelet consumption during renal replacement therapy – recorded in this study only at admission -, shock or bleeding). Missing data for the DIC-2001-72h and DIC-2018-72h, respectively, was another limitation. The reason for the missing data was two-fold: (1) ten patients died within the first 72 hours; (2) thirty-six patients were transferred before 72 hours and not all parameters used for the DIC score (e.g. D-dimer levels) are part of the routine laboratory work up for patients in other departments. Nevertheless, we were able to collect 80.8% of the data. As this study only included patients from an ICU with a cardiovascular focus, future studies should investigate the DIC score in patients from ICUs with a different focus. However, given the small sample size, future studies are needed to validate the utility in non-septic patients.
      In summary, the ISTH DIC score was predictive of mortality in non-septic patients, indicating a potential role for the ISTH DIC score in a broader population of non-septic ICU patients. Repeated assessment of the DIC score may be worthwhile and deserves further investigation.

      Authors’ contributions

      K.A. Krychtiuk and W.S. Speidl conceptualized the study and lead the investigation. B. Jilma incepted the analysis of repeated DIC in non-septic ICU patients. J. Grafeneder and K.A. Krychtiuk wrote the original draft as co-first authors. Reviewing and editing was done by B. Jilma and W.S. Speidl. J. Grafeneder and B. Jilma performed the formal and statistical analysis. All authors edited and approved the final manuscript.

      Source of funding

      Funded by the Austrian Science Fund FWF_F 5404-B21 (Special Research Program, Cellular Mediators Linking Inflammation and Thrombosis).

      Declaration of Competing Interests

      The authors state that they have no relevant conflict of interest.

      Acknowledgement

      This work was supported by the Association for the Promotion of Research on Arteriosclerosis, Thrombosis, and Vascular Biology, the Ludwig Boltzmann Cluster for Cardiovascular Research, and the Society of Thrombosis and Haemostasis Research. We would also like to thank Sarah Ely for providing language help and proof reading.

      References

        • Yamakawa K.
        • et al.
        Recombinant human soluble thrombomodulin in severe sepsis: a systematic review and meta-analysis.
        J Thromb Haemost. 2015; 13: 508-519
        • Gando S.
        • Levi M.
        • Toh C.H.
        Disseminated intravascular coagulation.
        Nat Rev Dis Primers. 2016; 2: 16037
        • Toh C.H.
        • Hoots W.K.
        • S.S.C.o.D.I.C.o.t. ISTH
        The scoring system of the scientific and standardisation committee on disseminated intravascular coagulation of the international society on thrombosis and haemostasis: a 5-year overview.
        J Thromb Haemost. 2007; 5: 604-606
        • Suzuki K.
        • et al.
        A re-evaluation of the D-dimer cut-off value for making a diagnosis according to the ISTH overt-DIC diagnostic criteria: communication from the SSC of the ISTH.
        J Thromb Haemost. 2018; 16: 1442-1444
        • Taylor Jr., F.B.
        • et al.
        Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation.
        Thromb Haemost. 2001; 86: 1327-1330
        • Buchtele N.
        • et al.
        Added value of the DIC score and of D-dimer to predict outcome after successfully resuscitated out-of-hospital cardiac arrest.
        Eur J Intern Med. 2018; 57: 44-48
        • Demma L.J.
        • et al.
        Predicting mortality in patients with disseminated intravascular coagulation after cardiopulmonary bypass surgery by utilizing two scoring systems.
        Blood Coagul Fibrinolysis. 2019; 30: 11-16
        • Krychtiuk K.A.
        • et al.
        Monocyte subset distribution is associated with mortality in critically ill patients.
        Thromb Haemost. 2016; 116: 949-957
        • Krychtiuk K.A.
        • et al.
        Mitochondrial DNA and Toll-Like Receptor-9 Are Associated With Mortality in Critically Ill Patients.
        Crit Care Med. 2015; 43: 2633-2641
        • Schwameis M.
        • et al.
        Asphyxia by Drowning Induces Massive Bleeding Due To Hyperfibrinolytic Disseminated Intravascular Coagulation.
        Crit Care Med. 2015; 43: 2394-2402
        • Gando S.
        • et al.
        A multicenter, prospective validation study of the Japanese Association for Acute Medicine disseminated intravascular coagulation scoring system in patients with severe sepsis.
        Crit Care. 2013; 17: R111
        • Dindo D.
        • et al.
        Kinetics of D-dimer after general surgery.
        Blood Coagul Fibrinolysis. 2009; 20: 347-352
        • Sniecinski R.M.
        • Chandler W.L.
        Activation of the hemostatic system during cardiopulmonary bypass.
        Anesth Analg. 2011; 113: 1319-1333
        • Frost S.A.
        • et al.
        Severity of illness and risk of readmission to intensive care: a meta-analysis.
        Resuscitation. 2009; 80: 505-510
        • Schwameis M.
        • et al.
        Prognosis of overt disseminated intravascular coagulation in patients admitted to a medical emergency department.
        Eur J Emerg Med. 2017; 24: 340-346
        • Gando S.
        • et al.
        Role of disseminated intravascular coagulation in severe sepsis.
        Thromb Res. 2019; 178: 182-188