Impact of coagulopathy assessment with thromboelastography and thromboelastometry on transfusion requirements in critically ill cirrhosis with nonvariceal bleeding: A prospective observational study

ABSTRACT

Background:

Viscoelastic tests are now routinely used for coagulopathy correction in patients with cirrhosis. Thromboelastography (TEG^®) and rotational thromboelastometry (RoTEM®) are the most widely studied tests in this population. However, they have not been compared with each other in critically ill patients with liver disease presenting with nonvariceal bleed. Hence, we aimed to compare these tests for coagulopathy correction in patients with liver disease presenting with nonvariceal bleeding.

Methods:

Sixty adult patients with liver cirrhosis presented to the liver intensive care unit, presenting with a nonvariceal upper gastrointestinal (GI) bleed (diagnosed by doing upper GI endoscopy which revealed bleeding from a nonvariceal source) oral or nasal bleed were enrolled. The patients were allocated to the TEG^® group (Group T) or RoTEM^® group (Group R) depending on the immediate availability of the viscoelastic test. Coagulopathy correction was done in each group as per established protocols and the results were compared.

Results:

There was a significant difference in the fresh frozen plasma (FFP) transfusion between the groups. The TEG^® group received more FFP when compared to the RoTEM^® group (P = 0.001).

Conclusion:

RoTEM^®-based coagulopathy correction leads to lesser use of blood products with similar control of bleeding when compared to TEG, in critically ill patients with cirrhosis.

INTRODUCTION

Patients with chronic liver disease are now known to have a rebalanced hemostasis with defects in pro-hemostatic drivers being compensated by changes in anti-hemostatic drivers.[1] These patients do not have consistent defects in hemostasis, but a delicate balance of pro and anti-hemostatic factors. Furthermore, deficiency of hemostatic factors has not been associated with bleeding events.[2]

Viscoelastic tests such as thromboelastography (TEG^®) and Rotational thromboelastometry (RoTEM^®) assess the viscoelastic properties of noncentrifuged whole blood.[3] They have been used in patients with mild to advanced cirrhosis. Results have shown that the coagulopathy can range from normal to hypocoagulable state and is independent of the severity of cirrhosis.[4]

TEG^® has been studied to predict bleeding after invasive procedures in critically ill patients with cirrhosis.[5] These patients have consistent delayed clotting, weak thrombus strength, and impaired fibrinogen function. However, once the clot is formed, there is evidence of reduced clot lysis.[6]

RoTEM^® has been correlated with laboratory parameters to study coagulopathy in stable cirrhosis. The maximum clot firmness (MCF) and clot formation time (CFT) correlate well with the platelet count and activity of antithrombin, fibrinogen, and Factor II. These two parameters effectively distinguish stable cirrhosis from healthy individuals. The platelet count is the rate-limiting factor for clot firmness in stable cirrhosis.[7]

These viscoelastic tests have been compared with laboratory parameters in cirrhosis undergoing invasive procedures.[6,7] They have been found to predict coagulopathy better as compared to the laboratory tests during liver transplantation, with a reduction in the total intraoperative blood loss, reduced fresh frozen plasma (FFP) consumption, and cost-effectiveness.[2,4]

No prospective studies have compared these tests in terms of clinical results in critically ill patients with liver disease. Hence, we aimed to compare the efficacy of TEG^® and RoTEM^® guided hemostatic management in cirrhosis with abnormal coagulation tests, presenting with nonvariceal bleeding in the intensive care unit (ICU) of our institution.

METHODS

After obtaining the proper ethics approvals, a prospective observational study was conducted from June 5, 2021, to January 10, 2022, in the Department of Anesthesiology and Critical Care at the Institute of Liver and Biliary Sciences in New Delhi, India. Written informed consent was required from all study participants (or their legal surrogate), and the manuscript adheres to the STROBE guidelines.

Patients fulfilling the following inclusion criteria were eligible for participation in the study: Patients with liver cirrhosis of any etiology (diagnosis of cirrhosis was based on previous liver biopsy and/or a combination of clinical, biochemical, and imaging findings) admitted to the liver ICU, aged 18–75 years, presenting with a nonvariceal upper gastrointestinal (GI) bleed (diagnosed by doing upper GI endoscopy which revealed bleeding from a nonvariceal source) oral or nasal bleed and coagulopathy as assessed by laboratory tests of international normalized ratio (INR) >1.5 or platelet counts <50,000/µL.

Exclusion criteria for the study were the following

Pregnancy, those who died within the first 24 h of admission in the ICU, presence of variceal bleeding, patients receiving anticoagulants or anti-platelets at the time of admission in ICU or had been discontinued <7 days before admission, transfusion of blood products within 24 h before admission to the ICU, significant cardiopulmonary disease.

Patients with liver cirrhosis presenting to the liver ICU with oral, nasal, or nonvariceal upper GI bleed were screened for the study. After fulfilling the inclusion and exclusion criteria the patients were allocated to the TEG^® group (Group T) or RoTEM^® group (Group R) depending on the immediate availability of the viscoelastic test.

Patients in the TEG^® (T) group received blood component therapy as follows: FFP transfusion (10 mL/kg body weight) was administered if the R time (reaction time) was more than 10 min. Cryoprecipitate (5 pooled units) was transfused if the alpha angle was <45°. Patients were transfused with a single donor apheresis platelet unit (SDAP) which is equivalent to 6–8 pooled units of platelets if the maximum amplitude (MA) was <55 mm.[8]

Patients in the RoTEM^® group received blood component therapy as per the algorithm.[9] If the extrinsically activated thromboelastometric test (EXTEM) A10 was 40 mm or less (and/or CFT was greater than 130 s), it suggested either a platelet or fibrinogen deficiency in the patient. Simultaneously if If the FIBTEM A10 was less than 10 mm, fibrinogen replacement was done with 6 units of cryoprecipitate, however if the FIBTEM A10 was 10 mm or single donor apheresis platelets were transfused. Normal EXTEM CT was defined as 43–95 s based on the manufacturer’s recommendations and elevations above 95 s were transfused with 10 mL/kg of FFP.

The quantity of blood products transfused in both the groups was determined by the ideal body weight which was calculated using the Devine formula[10] (male ideal body weight = 50 kg + 2.3 kg per inch over 5 feet; female ideal body weight = 45.5 kg + 2.3 kg per inch over 5 feet). At the time of allocation, demographic details, baseline hemogram, renal and liver function tests, chest X-ray, model for end-stage liver disease (MELD) score, sequential organ failure assessment score, coagulation profile, INR, and baseline TEG^® or RoTEM^® variables were noted.

The TEG^® or RoTEM^® was repeated at 8 h to look for improvement in the viscoelastic coagulation parameters. The amount of blood products transfused was noted and the patients were monitored for any transfusion-related side effects. Hemoglobin was monitored every 6 h to assess for control of bleed. Patients were assessed at 48 h for control of bleeding after the coagulopathy correction.

Failure to control bleed was defined by the presence of one of the following: Fresh hematemesis, nasal, or oral bleed or nasogastric aspiration of 100 mL after 2 h of initial treatment, drop in hemoglobin by 2 g%, or development of hypovolemic shock.[11] The primary endpoint was the total quantity of blood products transfused at 8 h.

Secondary endpoints were the following: control of bleeding at 48 h, improvement in the TEG^®/RoTEM^® parameters at 8 h, transfusion-related side effects such as transfusion-related acute lung injury (TRALI) and transfusion-associated circulatory overload (TACO).

Sample size

Assuming that the number of products required would be 7.1 ± 2.3 in the TEG^® group and 5.43 ± 1.8 in the RoTEM^® group, with a margin of error of 5% and power of 90, we needed to enroll 66 cases. A total of 60 patients were enrolled for the study with 30 patients allocated in each arm.

Assessment techniques

Thromboelastometry^®

RoTEM^® DELTA (TEM Systems, Inc., Durham, NC) was used for viscoelastic assessment. Only EXTEM and FIBTEM analysis was performed in our institution using the appropriate reagents provided by the manufacturer. In the EXTEM test, tissue factor was added as a reagent to activate the coagulation and in the FIBTEM test, a platelet inhibitor (cytochalasin D) was added for the evaluation of the fibrinogen function. The following parameters were measured for the hemostatic analysis: CT: it was defined as the time from the start of the measurement to the initiation of clotting (Clot firmness 2 mm); CFT: it was defined as the time from the initiation of clotting till the clot reached a firmness of 20 mm; alpha angle: defined as the tangent to graphical trace at an amplitude of 2 mm; A 10: amplitude after 10 min of CT; MCF: MA (mm) of the graphical trace of clot firmness; and ML30: Maximum lysis at 30 min.

Thromboelastography^®

A kaolin-activated TEG^® assay was performed with a 5000 series (Haemoscope, Inc., Niles, IL). The specific TEG^® variables used in the study to guide blood component transfusions were: R time: defined as the time from the start of the test to initial fibrin formation (amplitude of 2 mm); K time: defined as the time taken to achieve a amplitude of 20 mm; alpha angle: tangent to graphical trace; MA: MA of the graphical trace of clot firmness; and LY30: lysis at 30 min.

Transfusion-related events

For the purposes of this study, “serious” transfusion reactions included cardiopulmonary, hemolytic, septic, hypotensive, or anaphylactic reactions; and “minor” transfusion reactions included febrile nonhemolytic and minor allergic reactions.

Transfusion-related acute lung injury

TRALI was defined by the following criteria:[12] acute onset, hypoxemia (P/F ratio <300 or SpO₂ <90% on room air), clear evidence of bilateral pulmonary edema on imaging, onset within 6 h of transfusion, no temporal relationship to an alternative risk factor for acute respiratory distress syndrome.

Transfusion-associated circulatory overload

TACO was defined by the following criteria:[13] new onset or exacerbation of respiratory symptoms within 6 h of transfusion cessation with inclusion of three or more of the following: acute respiratory distress, manifested by cough, dyspnea, orthopnea, and tachypnea; elevated brain natriuretic peptide; elevated central venous pressure; evidence of left heart failure; evidence of positive fluid balance; and/or radiographic evidence of pulmonary edema.

Statistical analysis

Data analysis was done using the Statistical Package for Social Sciences (SPSS) for Windows, Version 22.0 software (IBM Corp.; Armonk, NY, USA). Continuous data were analyzed using the Student’s t-test and the Mann–Whitney test. Categorical data were analyzed using the Chi-square and Fischer’s test. The change of measures over time was analyzed by repeated measure analysis. A P < 0.01 was considered statistically significant.

RESULTS

A total of 185 patients with liver cirrhosis presented in our hospital with upper GI bleeding. Out of these, 102 patients had variceal bleeding and 83 patients had nonvariceal bleeding. Five patients with nonvariceal bleeding who were transfused with blood products died within 6 h of transfusion and were excluded from the study. After the fulfillment of all the inclusion and exclusion criteria, 60 patients were enrolled in the study. 30 patients were allocated to the TEG^® group (Group T) and 30 patients were allocated to the RoTEM^® group (Group R). The study was stopped after the attainment of the required sample size. No significant differences were seen between the groups in terms of age, sex, liver disease scores, clotting variables, and the clinical presentation of the patients [Table 1]. Seven patients (23%) in Group T and four patients (13%) in Group R had low platelets (<50,000) (P = 0.31). Twenty-five patients (83%) in Group T and 29 patients (96%) in Group R had deranged INR (>1.8) (P = 0.1).

Table 1.

Baseline demographic, clinical, and biochemical characteristics

Characteristic	TEG	ROTEM	P
Age	55.8±11.21	51.67±12.47	0.232
Female, n (%)	7 (23.33)	7 (23.33)	0.88
Male, n (%)	23 (76.66)	23 (76.66)
Clinical presentation, n (%)
Oral bleed	17 (56.67)	19 (63.33)	0.598
Nasal bleed	0	2 (6.67)	0.15
Malena	21 (70)	17 (56.67)	0.284
Hemoglobin	7.2 (7.025–7.2)	7.3 (6.9–7.7)	0.234
Total leucocyte count	14,580 (12,150–16,416)	14,800 (12,560–17,650)	0.403
Total bilirubin	11 (8.3–14.2)	10.43 (8.67–15.30)	0.371
AST	149.5±24.5	134.4±45.4	0.113
ALT	66±13.9	62.1±18.1	0.357
ALP	69.7±20.4	77.5±15.7	0.099
Albumin	3.24 (2.87–3.65)	3.42 (2.63–3.75)	0.455
Urea	119±39.2	136±49.8	0.149
Creatinine	2.1 (1.8–2.3)	1.85 (1.34–2.30)	0.343
Platelets	86,000 (49,500–140,000)	100,000 (69,000–154,000)	0.11
INR	2.4 (2–3.5)	2.75 (2.4–3.4)	0.159
Lactate	1.9 (1.6–3.35)	2.4 (2–3.5)	0.155
MELD score	25.1±4.56	25.0±5.57	0.955
Fibrinogen	49.8±4.69	49.7±4.94	0.915

TEG: Thromboelastography, ROTEM: Rotational thromboelastometry, AST: Aspartate aminotransferase, ALT: Alanine aminotransferase, ALP: Alkaline phosphatase, INR: International normalized ratio, MELD: Model for end-stage liver disease

Nineteen patients (63.33%) presented with oral bleeding while 17 patients (56.67%) presented with melena in the group R. Seventeen patients (56.67%) presented with oral bleeding and 21 patients (70%) had melena in the group T. Only two patients had nasal bleeding who underwent RoTEM^® for the correction of the coagulopathy.

Control of bleed

At the end of 48 h, control of bleeding was seen in 16 (53.3%) patients in the group T and 20 (66.6%) patients in the group R [Table 2].

Table 2.

Control of bleeding in both groups at 48 h

Variable	TEG	ROTEM	P
Bleeding	14	10	0.292
No bleeding	16	20

TEG: Thromboelastography, ROTEM: Rotational thromboelastometry

Blood products transfused and complications

There was a significant difference in the FFP transfusion with group T receiving more FFP when compared to group R [P = 0.001; Table 3]. No significant difference was seen in the packed red blood cell, cryoprecipitate, single donor platelet, and pooled donor platelets (PDP) transfusion in both the groups. There was no significant difference between the two groups in the incidence of transfusion-associated acute lung injury and transfusion-associated with circulatory overload [Table 4].

Table 3.

Blood products transfused in both groups

Variable (units transfused per patient)	TEG	ROTEM	P
Packed RBC	0.67±0.8	0.37±0.81	0.057
FFP	3.0±1.46	0.8±1.63	0.001
Cryoprecipitate	1.2±2.44	2.6±3.02	0.054
Single donor platelets	0.13±0.35	0.23±0.43	0.321

RBC: Red blood cell, FFP: Fresh frozen plasma, TEG: Thromboelastography, ROTEM: Rotational thromboelastometry

Table 4.

Complications of transfusion

	TEG, n (%)	ROTEM, n (%)	P
TRALI absent	27 (90)	28 (93.3)	0.64
TRALI present	3 (10)	2 (6.66)
No TACO	28 (93.3)	29 (96.66)	0.554
TACO	2 (6.66)	1 (3.33)

TRALI: Transfusion-related acute lung injury, TACO: Transfusion-associated circulatory overload, TEG: Thromboelastography, ROTEM: Rotational thromboelastometry

The mean value of R time at admission was (16.12 ± 4.71) which significantly reduced to (9.83 ± 3.86) after transfusion (P = 0.001). Similarly, there was a significant improvement in MA from (57.8 ± 14.52) mm to (64.3 ± 11.35) mm after platelet transfusion (P = 0.018). However, there was no significant change in K time, alpha angle, and lysis index after correction [Table 5].

Table 5.

Change in TEG variables after transfusion. (Wilcoxon signed rank test)

Variable	0 h	8 h	P
Reaction time (R)	16.12±4.71	9.83±3.86	0.001
K time	5.36±1.73	4.73±1.62	0.046
Alpha angle	53.53±14.33	54.36±5.99	0.718
MA	57.8±14.52	64.3±11.35	0.018
LY30 index	0.86±2.44	0.51±1.80	0.109

MA: Maximum amplitude, LY30: Lysis 30

In group R, patients underwent EXTEM and FIBTEM. The EXTEM CT significantly improved from (70.9 ± 14.6) to (66.63 ± 12.91) after correction (P = 0.002). The EXTEM CT and MCF both showed a significant improvement after correction (P = 0.001). However, the A10 and Lysis 30 did not show a significant change [Table 6]. In contrast, all the parameters of FIBTEM CT, MCF, and A10 showed significant improvement after correction (P < 0.001). Similarly, the Lysis 30 showed significant improvement after correction (P = 0.009).

Table 6.

Change in rotational thromboelastometry variables after transfusion

Variable	0 h	8 h	P
EXTEM CT	70.9±14.6	66.63±12.91	0.002
EXTEM CFT	185.9±23.11	145.93±18.53	0.001
EXTEM A10	54.3±11.8	51.67±9.01	0.054
EXTEM MCF	47.63±11.13	59.73±7.32	0.001
EXTEM LI 30	96.5±5.51	98.3±2.9	0.047
FIBTEM CT	96.27±13.98	84.17±17	0.001
FIBTEM MCF	13.97±6.20	10.47±3.57	0.001
FIBTEM A10	13.67±4.53	10.3±2.93	0.001
FIBTEM LI 30	102.68±6.35	99.90±3.30	0.009

EXTEM: Extrinsically activated thromboelastometric test, FIBTEM: Fibrin-based thromboelastometric test, ROTEM: Rotational thromboelastometry, CT: Clotting time, CFT: Clot formation time, A10: Alpha10, MCF: Maximum clot formation, LI 30: Lysis index 30

DISCUSSION

Cirrhosis is characterized by decreased synthesis of both procoagulants and anticoagulants, whose delicate balance is further weakened by thrombocytopenia and/or thrombocytopathy.[14] These abnormalities result in the prolongation of prothrombin time and activated partial thromboplastin time, all of which have led to cirrhosis being considered a prototypical hemorrhagic disorder. INR is no longer considered a reliable test for predicting bleeding in patients with liver cirrhosis.[15] Viscoelastic tests such as TEG^® and RoTEM^® have been enthusiastically proposed and proven to be superior to traditional lab tests. Though not considered a gold standard, they have shown benefit in guiding transfusion-based decisions in elective cardiac procedures and liver transplantation.[16] Besides laboratory tests have a longer turnaround time, which presents a challenge in rapidly bleeding patients. Therefore, point-of-care viscoelastic tests overcome these disadvantages and provide a rapid, global, and functional assessment of the clotting system. This results in better patient management and improved outcomes.

The goal in bleeding patients with cirrhosis is to use appropriate blood products, reduce the risk of volume overload, transfusion reactions, blood sensitization, and other complications of blood product infusion.[17]

There are no clear guidelines in literature regarding the management of coagulopathy in patients with liver cirrhosis with nonvariceal bleeding. The aim of our study was to compare a TEG^® and RoTEM^®-guided blood transfusion strategy in terms of transfusion requirements and outcomes.

In our study, an increased number of FFPs were transfused in the TEG^® group when compared to the RoTEM^® group (P < 0.001). TEG^® based algorithms are known to favor the use of FFP.[18] Furthermore, RoTEM^® system seems to have shorter reaction times, steeper angles, and higher MA, resulting in the impression that the patient is less coagulopathic.[19] A difference in FFP use in both the groups may be a consequence of these differences. Our study showed that there was a significant improvement in the FIBTEM CT, FIBTEM A10, FIBTEM MCF, FIBTEM LI30, EXTEM CT, EXTEM CFT, EXTEM MCF, EXTEM LY30 values after correction of coagulopathy. It also demonstrates statistically significant improvement in the R time and K value and MA value of the TEG^® measurements.

However, there was no difference in the number of platelets, packed RBC, cryoprecipitate transfused between the TEG^® and the RoTEM^® groups.

In our study, 10% of the patients developed TRALI in the TEG^® group compared to 6.6% of patients in the RoTEM^® group (P = 0.64). A study by Benson et al. in patients admitted to intensive care with GI bleeding showed the incidence of TRALI to be 15% with an increased incidence of 29% in patients of end-stage liver disease. This study showed a temporal association with FFP transfusion in these patients with each unit of FFP increasing the incidence of TRALI by 11%.[20] Serum albumin and MELD scores were found to be independent nontransfusion risk factors for a high incidence of TRALI. However, this study was a retrospective study, with no transfusion protocol. Ours is a prospective study with a uniform transfusion protocol for blood product transfusion.

In our study, there was no significant difference in the incidence of TRALI between the two groups, despite more FFP transfusion in the TEG^® group (P < 0.001). This could be attributed to the small sample size used in our study.

A retrospective study performed in patients with GI bleeding found the incidence of TACO to be 12.3% with increased chances of TACO development in patients with advanced-stage of liver disease.[21] However, in our study, 6.6% of patients developed TACO in the TEG^® group as compared to 3.3% in the RoTEM^® group (P = 0.55). There is a paucity of studies about the risk factors and incidence of TACO in this subset of patients and more prospective studies are required in the future to study the true incidence of TACO in this subset of patients.

Follow-up of bleeding at 48 h between the two groups was insignificant (P = 0.292). 10 patients among the RoTEM^® group and 14 among the TEG^® group had a failure of control of bleeding.

At our institution, we have adopted viscoelastic testing as the standard of care, gradually replacing the use of platelet count or INR to guide blood transfusion for preprocedure prophylaxis for invasive procedures or management of coagulopathy-related GI bleeding. Our results are consistent with previous studies that separately evaluated TEG^® and RoTEM^® in cardiac surgery.[22,23,24] A few studies in literature have compared these viscoelastic tests in patients of trauma.[25,26,27] Hagemo et al.[26] have demonstrated that inter-changeability between TEG^® and RoTEM^® is limited in the trauma patients and the development of separate treatment algorithms is required for the two devices. Rizoli et al.[27] have demonstrated that these viscoelastic tests are similar in clinical applicability but are not interchangeable due to different coagulation triggers.

There are no studies in literature comparing TEG^® versus RoTEM^®-guided blood transfusion in a liver intensive care setup, while there are studies comparing the repeatability of TEG^® in stable cirrhosis.[28] A study comparing TEG^® and RoTEM^® in liver transplant recipients found the values in TEG^® and RoTEM^® are not interchangeable which could be attributed to the different coagulation activators used in the devices.[29]

TEG^®/RoTEM^®-guided transfusion strategy has been found to be associated with a significantly lower use of blood components compared with transfusion guided by INR and platelet count, without an increase in bleeding complications, in patients with cirrhosis undergoing invasive procedures.[30,31] Additional studies are needed to identify the predictive values for a range of trigger points based on TEG^® and RoTEM^® parameters. It remains to be evaluated whether the threshold TEG^®/RoTEM^® parameters (trigger points) we chose for the transfusion of blood components could be further relaxed to pick up patients with more severe coagulopathy, and thus further reduction in the need for blood component transfusion.

There are some limitations of our study. The small sample size in our study limits the generalization of our results and further prospective studies are needed recruiting a large number of patients. Our study was a prospective study in which TEG^® or RoTEM^® guided transfusion was done depending on the immediate availability of the test. It was not a randomized study which could have led to a bias in the study. We only followed up with the patients for 48 h after coagulopathy correction as per the design of the study and did not monitor for the levels of individual clotting factors and platelet function tests before and after blood component transfusion in our study.

CONCLUSION

Patients with advanced cirrhosis with coagulopathy with nonvariceal upper GI bleeding, RoTEM^®-based coagulopathy correction led to lesser use of blood products with similar control of bleeding as compared to TEG^®-based correction. Further prospective studies are required to evaluate if the thresholds used for coagulopathy correction in our study can be reduced further to decrease the amount of blood product transfusion in this subset of patients.

Research quality and ethics statement

This study was approved by the Institutional Review Board/Ethics Committee at Institute of Liver Biliary Sciences (Approval #: IEC/2021/86/MA06; Approval date: June 3, 2021). The authors followed the applicable EQUATOR Network (http://www.equator-network.org/) guidelines, specifically the STROBE Guideline, during the conduct of this research project.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.