Introduction
Traditional coagulation studies (especially the INR) fail miserably in cirrhosis. Thromboelastography (TEG) is a superior approach for understanding the global balance of pro-coagulants versus anti-coagulants in these patients. This isn’t anything particularly new – for example, it was explored in this post from 2015 (if you're not familiar with this concept already, it's explained in that post).
Unfortunately, implementing the use of TEG in cirrhosis has been difficult for several reasons:
- The precise strategy for using TEG is unclear. Exactly how should TEG be used to direct blood product transfusion?
- Lack of prospective data demonstrating that TEG is superior to conventional coagulation tests for non-surgical bleeding.
- Many physicians with internal medicine training aren’t comfortable with TEG.
Thus, a strange double-standard has existed regarding the use of TEG to guide product administration in cirrhosis. In the operating room, TEG is used, validated, and generally accepted. Meanwhile, in the ICU, TEG has remained under-utilized. These disparities don’t make sense (coagulation is coagulation).
This post will review some RCTs comparing TEG with conventional coagulation tests. This is a rare treat! The vast majority of laboratory studies are supported by exactly zero randomized controlled trials (e.g. checking INR levels in cirrhosis is supported by – you guessed it – zero RCTs). Rather than fixing numbers just because that seems right, an RCT allows us to test whether this is actually helping patients.
four RCTs comparing conventional coagulation testing with TEG
Wang et al 2010: Thromboelastography-Guided Transfusion Decreases Intraoperative Blood Transfusion During Orthotopic Liver Transplantation: Randomized Clinical Trial
This is a single-center RCT involving 28 patients undergoing liver transplantation.1 Patients were randomized into two groups, with coagulation strategies as follows:
Conventional coagulation test group:
- Platelets <50 billion/L ➡️ Platelet transfusion
- Fibrinogen <100 mg/dL ➡️ Cryoprecipitate transfusion
- INR >1.5 times control ➡️ Fresh frozen plasma transfusion
TEG group:
- Coagulation was monitored using a kaolin-activated TEG
- R-time prolonged (>10 minutes) ➡️ Fresh frozen plasma transfusion
- MA reduced (<55 mm) ➡️ Platelet transfusion
- Alpha-angle reduced (<45 degrees) ➡️ Cryoprecipitate transfusion
Patients in the TEG group received half as much fresh frozen plasma on average (table below). There were trends towards less blood loss in the TEG group.
This study demonstrates that a TEG-based transfusion strategy can be safe and potentially beneficial in a group of patients undergoing major hemostatic challenge (liver surgery). It implies that a conventional coagulation test strategy will lead to over-utilization of blood products. However, its statistical power is limited by the small size.
Kumar et al 2019: Thromboelastography-Guided Blood Component Use in Patients with Cirrhosis with Nonvariceal Bleeding: A Randomized Controlled Trial
This is a randomized controlled trial involving 96 patients with cirrhosis and non-variceal gastrointestinal bleeding.2 Key inclusion criteria were:
- Status post upper endoscopy showing non-variceal GI bleed with ongoing bleeding
- Significant coagulopathy (defined as INR >1.8 and/or platelets <50 billion/L)
- Not on anti-platelet or anticoagulant therapy
- Not undergoing chronic hemodialysis
Patients were randomized to receive blood products based on conventional coagulation testing or TEG, according to the following algorithms:
Conventional coagulation testing arm:
- INR >1.8 ➡️ Fresh frozen plasma transfusion.
- Platelet count <50 billion/L ➡️ Platelet transfusion.
- Fibrinogen level <80 mg/dL ➡️ Cryoprecipitate transfusion.
TEG arm (this is identical to Wang et al above):1
- Coagulation was monitored using a kaolin-activated TEG
- Elevated R-time (>10 minutes) ➡️ Fresh frozen plasma transfusion.
- MA reduced (<55 mm) ➡️ Platelet transfusion.
- Alpha-angle reduced (<45 degrees) ➡️ Cryoprecipitate transfusion.
Patient groups were well-matched at baseline and generally had severe coagulopathy. The following are the baseline characteristics of the overall patient group:
- Enzymatic coagulation
- INR: prolonged >1.8 in 90% of patients
- R-time: prolonged >10 minutes in 60% of patients
- Platelets function
- Platelet count: <50 billion/L in 80% of patients
- Maximal Amplitude: <55 mm in 53% of patients
- Fibrinogen function
- Fibrinogen: <80 mg/dL in 95% of patients
- Alpha-angle: <45 degrees in 60% of patients
It is immediately obvious from these baseline characteristics that patients in the TEG arm will receive fewer blood products. And this was indeed demonstrated: patients in the TEG arm received fewer of every type of blood product (table below). Some patients in the TEG arm managed to escape receiving any coagulation product, whereas this never happened in the conventional coagulation test arm:
As shown above, patients in the TEG arm were significantly less likely to experience any transfusion-related complications (p<0.001). This comes as no surprise, as it is nearly guaranteed by a more restrictive transfusion strategy.
What about hemostasis? There was no difference in the required amount of RBC transfusions. However, patients in the TEG arm experienced recurrent bleeding less frequently and spent less time in the ICU:
This is the largest RCT comparing TEG to conventional coagulation tests. It implies that a traditional transfusion strategy leads to over-transfusion and more adverse reactions from blood products, while simultaneously leading to worse hemostasis! How can we understand this riddle? Many patients in this study had portal hypertensive gastropathy. Over-transfusion will increase portal pressures, increasing the pressure head driving this bleed.
Rout et al 2019: Thromboelastography-guided blood product transfusion in cirrhosis patients with variceal bleeding
This study took TEG a step farther, looking at it in variceal bleeding – generally the most severe form of bleeding encountered in cirrhosis.3 Key inclusion criteria were:
- Cirrhosis with acute variceal bleeding
- Significant coagulopathy (defined as platelets <50,000 and/or INR > 1.8)
- No use of anti-platelet agents or NOACs
- Not in acute or chronic renal failure
- Not in shock
- Not septic
Patients were randomized to receive blood products based on conventional coagulation testing or TEG according to the following algorithm. This algorithm is different from the above algorithm. One difference is that they used native TEG (rather than kaolin-activated TEG, which is faster and thus generally used to guide transfusion in hemorrhage).
Conventional coagulation testing arm:
- INR elevation >1.8 ➡️ Fresh frozen plasma administration.
- Platelets <50,000 ➡️ Platelet transfusion.
Native-TEG arm (cutoffs were based on values outside the normal range):4
- Elevated R-time (>15 minutes) ➡️ Fresh frozen plasma administration.
- Reduced MA (<30 mm) ➡️ Platelet transfusion
Patients randomized to the TEG arm received fewer blood products and had a decreased risk of re-bleeding. There was a trend towards reduced mortality among the TEG group, but the study was underpowered to prove this. This result mirrors the results of the last study by Kumar et al:2 TEG-based transfusion involved administration of fewer blood products, while achieving better hemostasis.
One interesting aspect of this study is that no cryoprecipitate was used and no fibrinogen levels were checked! Despite this, patients in the TEG group did very well. How is it possible to get by without cryoprecipitate? More on this riddle below.
De Pietri 2016: Thromboelastography-Guided Blood Product Use Before Invasive Procedures in Cirrhosis with Severe Coagulopathy: A Randomized, Controlled Trial.
This was a study of using TEG to guide blood product administration prior to a variety of procedures (as shown above).5 Key inclusion criteria were:
- Significant coagulopathy (defined as INR > 1.8 and/or platelets <50)
- No active bleeding
- Not on anti-platelet or anticoagulant medication
- Not on hemodialysis
- Not septic
Sixty patients were randomized to a conventional coagulation testing arm or a TEG arm. Like Rout et al 2019,3 this study used the slower native TEG assay. Treatment of the two groups was as follows:
Conventional coagulation testing arm:
- INR > 1.8 ➡️ Fresh frozen plasma administration.
- Platelets <50 ➡️ Platelet administration.
Native-TEG arm:
- Substantially elevated R-time (>40 minutes, compared to their normal range of 12-26 minutes) ➡️ Fresh frozen plasma administration.
- Substantially reduced MA (<30 mm, compared to their normal range of 42-63 mm) ➡️ Platelet transfusion
All patients in the conventional coagulation test arm received blood products, whereas 93% of patients in the TEG arm didn’t receive any blood product at all (table below). Only one patient had a bleeding complication (a patient in the standard coagulation test group who received a large-volume paracentesis).
This study illustrates again that conventional coagulation tests lead to over-utilization of blood products. Unfortunately, this study wasn’t powered to evaluate patient-centered outcomes (e.g. adverse events from blood transfusion).
reconciling different TEG-based transfusion algorithms
The four studies above are split between two TEG-based strategies.
Wang et al. & Kumar et al.:1,2
- Prolonged R-time ➡️ Give FFP
- Low MA ➡️ Give platelets
- Low alpha-angle ➡️ Give fibrinogen
Rout et al. & De Pietri et al.:3,5
- Prolonged R-time ➡️ Give FFP
- Low MA ➡️ Give platelets
- [Don’t give fibrinogen]
Neither of these strategies is perfect. Ignoring fibrinogen entirely seems unwise for patients with cirrhosis and active bleeding (especially if they are receiving massive blood transfusion). These patients may have low fibrinogen at baseline, or their fibrinogen may be diluted by other blood products. Fibrinogen is one of the key materials that clots are physically made of, so a severe deficiency will impair coagulation.
However, using alpha-angle as a transfusion target to give fibrinogen isn’t perfect either. Although alpha-angle is widely believed to measure fibrinogen levels, this is a myth.6 Alpha-angle is influenced by both fibrinogen and platelet function. In patients with a low alpha-angle due to thrombocytopenia, administration of fibrinogen may theoretically be unnecessary.
More sophisticated TEG-based transfusion algorithms are available to parse out exactly who might benefit most from fibrinogen versus who might benefit most from platelets. Specifically, TEG may be modified with platelet inhibitors to obtain a functional fibrinogen assay, which does accurately measure fibrinogen levels. Another study by De Pietri et al. described the effects of implementing a sophisticated TEG-based transfusion algorithm incorporating functional fibrinogen (compared to a TEG-based algorithm that ignored fibrinogen entirely). Fibrinogen administration spared other blood products, without affecting mortality.7
This pre-post study on fibrinogen illustrates an interesting paradox. Patients seemed to benefit from fibrinogen (with less bleeding). However, patients also fared OK with zero fibrinogen (experiencing an equivalent mortality). Why? Fibrinogen and platelets together form a physical clot, as shown in the equation below. The rate of this reaction depends on the levels of both factors. Thus, to some extent, a deficiency in one component can be compensated for by an elevated level of the other. This explains why both platelets and fibrinogen affect both alpha-angle and the MA – they’re working together in a synergistic function. This also implies that if you give a bit too much of one and not quite enough of another, you’ll be OK – they will work together to get the job done.
[Fibrinogen] + [Platelets] ➡️ Clot
Reaction rate = k[Fibrinogen][Platelets]
So ultimately the precise amount of fibrinogen or platelets that we use may not matter critically. The key benefit from a TEG-based strategy in cirrhosis likely stems from the use of R-time to avoid bombarding these patients with fresh frozen plasma.
Which TEG algorithm should we use?
This is a bit frustrating. From a theoretical standpoint, the best TEG algorithm would be one which explicitly measures fibrinogen (e.g. using a functional fibrinogen assay). However, there are several drawbacks to this more sophisticated strategy:
- Most hospitals don’t have access to a functional fibrinogen TEG assay.
- A more complex algorithm may be harder to implement (e.g. running several TEG assays could delay therapy).
- More sophisticated TEG algorithms haven’t been validated in prospective RCTs in the context of ICU patients with cirrhosis. Although such algorithms make sense, we can’t necessarily assume that this would translate into clinical benefit.
Ultimately, at this point in time, an algorithm using standard TEG may be the best solution for most units. This is the best validated approach (supported by RCTs in both in liver transplant and in gastrointestinal hemorrhage).1,2 Standard TEG assays are more widely available, especially in ICUs which don’t generally have on-site TEG. The algorithm is easy to use and understand. Ultimately, any TEG-based algorithm is likely to be a considerable improvement over conventional coagulation tests, with the precise algorithm being less important as shown below (I know you want a Tesla, but what you actually need is a Honda).
algorithm
Low-risk procedures (e.g. paracentesis, central line) – These can generally be done without any blood product administration. The key is performance by an expert operator under ultrasound guidance. For diagnostic paracentesis, identify the inferior epigastric artery with ultrasound and be sure to avoid it. For central line, stay far away from any arteries. If bleeding occurs, then give blood products as necessary.
Moderate-risk procedures (e.g. endoscopy with biopsy, liver biopsy) – Prior to the procedure, consider administration of FFP or platelets if the R-time or MA are considerably abnormal.5,8 Follow the patient carefully and give additional blood products if bleeding should occur (or if bleeding is visualized during the procedure itself).
Active hemorrhage or severe hemostatic stress – This may be managed using a standard TEG-based transfusion algorithm, as discussed above.1,2
Limitations & additional considerations
- These algorithms don’t account for medication-induced coagulopathy (e.g. clopidogrel, aspirin, or NOACs). Additional blood products or medications may be required for management of such coagulopathies.
- For patients with renal failure, uremia may cause platelet dysfunction. The administration of DDAVP may be considered for such patients in the face of active hemorrhage.
- Hyper-fibrinolysis isn’t uncommon in severe liver disease (the failing liver doesn't metabolize endogenous tPA, which causes patients to behave as if they were on a continuous tPA infusion). Unfortunately, standard TEG assays aren’t terribly sensitive for this. Hyperfibrinolysis may manifest clinically with persistent bleeding and inability to maintain adequate fibrinogen levels. Measurement of fibrinogen levels may be useful to diagnose this. If hyperfibrinolysis is proven or suspected, tranexamic acid may be useful (an ongoing infusion of high-dose tranexamic acid may be required to achieve hemostasis).
What about ROTEM? ROTEM and TEG are very similar viscoelastic tests. TEG predominates in the United States, whereas ROTEM is more popular in Europe. This post focuses on TEG because this is the methodology used in the RCTs and available locally. However, these principles should be easily adaptable to folks using the ROTEM system.
- Problems with standard coagulation tests in cirrhosis (especially INR) are well known.
- Four RCTs show benefit of blood product administration guided by thromboelastography (TEG), compared to traditional coagulation tests. These are all single-center studies, but they seem to give a coherent message: use of traditional coagulation tests leads to over-transfusion.
- The optimal TEG-based transfusion algorithm is unclear. The studies above showed benefit from two very different TEG-based strategies. It's likely that any reasonable TEG-based strategy will represent an improvement over strategies using traditional coagulation tests.
- An approach to TEG-based transfusion in cirrhosis is provided, which is based upon these RCTs:
Related
References
- 1.Wang S, Shieh J, Chang K, et al. Thromboelastography-guided transfusion decreases intraoperative blood transfusion during orthotopic liver transplantation: randomized clinical trial. Transplant Proc. 2010;42(7):2590-2593. https://www.ncbi.nlm.nih.gov/pubmed/20832550.
- 2.Kumar M, Ahmad J, Maiwall R, et al. Thromboelastography-Guided Blood Component Use in Patients With Cirrhosis With Nonvariceal Bleeding: A Randomized Controlled Trial. Hepatology. May 2019. https://www.ncbi.nlm.nih.gov/pubmed/31148204.
- 3.Rout G, Shalimar, Gunjan D, et al. Thromboelastography-guided Blood Product Transfusion in Cirrhosis Patients With Variceal Bleeding: A Randomized Controlled Trial. J Clin Gastroenterol. April 2019. https://www.ncbi.nlm.nih.gov/pubmed/31008867.
- 4.Subramanian A, Albert V, Saxena R, Agrawal D, Pandey R. Establishing a normal reference range for thromboelastography in North Indian healthy volunteers. Indian J Pathol Microbiol. 2014;57(1):43-50. https://www.ncbi.nlm.nih.gov/pubmed/24739830.
- 5.De P, Bianchini M, Montalti R, et al. Thrombelastography-guided blood product use before invasive procedures in cirrhosis with severe coagulopathy: A randomized, controlled trial. Hepatology. 2016;63(2):566-573. https://www.ncbi.nlm.nih.gov/pubmed/26340411.
- 6.Solomon C, Schöchl H, Ranucci M, Schlimp C. Can the Viscoelastic Parameter α-Angle Distinguish Fibrinogen from Platelet Deficiency and Guide Fibrinogen Supplementation? Anesth Analg. 2015;121(2):289-301. https://www.ncbi.nlm.nih.gov/pubmed/26197367.
- 7.De P, Ragusa F, Deleuterio A, Begliomini B, Serra V. Reduced Transfusion During OLT by POC Coagulation Management and TEG Functional Fibrinogen: A Retrospective Observational Study. Transplant Direct. 2015;2(1):e49. https://www.ncbi.nlm.nih.gov/pubmed/27500243.
- 8.Wang S, Lin H, Chang K, et al. Use of higher thromboelastogram transfusion values is not associated with greater blood loss in liver transplant surgery. Liver Transpl. 2012;18(10):1254-1258. https://www.ncbi.nlm.nih.gov/pubmed/22730210.
Image credit: High-wheel bicycles
- PulmCrit Blogitorial – Use of ECGs for management of (sub)massive PE - March 24, 2024
- PulmCrit Wee: Propofol induced eyelid opening apraxia – the struggle is real - March 20, 2024
- PulmCrit wee: Why I like central lines for GI bleed resuscitation - March 13, 2024
Fantastic post. Would you please do the same post for DIC?
Awesome post ,much needed very crisp and concise
Cheers
where I can find the cast(the audio format for this episode i’m totally confuse!! please help, thank you
Great post, really interested in TEG and how it applies to liver pt. We see often in house for ECMO and other heart/ICU pt, hadn’t seen it extended to liver pt yet though. I’ll be interested to see if it supplants PT/INR in any of the predictive models in the future