There are many overused tests in the vastness of medicine, but few synthesize both overuse and inappropriateness like the international normalized ratio (INR). The INR is a test with a singular purpose: a standardized calibration for the measurement of the anticoagulant activity of warfarin. It has no other reason for existence, yet it is frequently ordered as part of routine screening for previously undiagnosed coagulopathy. Indeed, its use is so prevalent as a stowaway in anachronistic order sets that ACEP once considered tracking coagulation studies as a marker of low-quality care.¹ The INR and its related cousin prothrombin time (PT) are elevated only in a handful of conditions: vitamin K–dependent factor deficiency, presence of factor VII inhibitor, disseminated intravascular coagulation, or in the setting of massive transfusion. An incidentally encountered mild elevation of INR or PT is typically of no clinical significance nor is it associated with increased bleeding risk.

The prime example of this misunderstanding and misuse of the INR is in our approach to patients with cirrhosis or end-stage liver disease (ESLD). As a result of their chronic liver injury, these patients have impaired intrahepatic synthesis of a number of important elements of hemostasis. Further complicating the matter, much of our contact with patients suffering from ESLD occurs in the heat of critical illness such as upper gastrointestinal bleeding, including brisk variceal bleeding. This is, unfortunately, where the INR does us the greatest disservice.

The INR does not accurately reflect underlying coagulopathy. Patients with impaired hepatic synthetic dysfunction and ESLD do suffer from a diminished hemostatic reserve (ie, thrombocytopenia, low levels of coagulation factors, and dysfibrinogenemia).² However, these deficiencies are counterbalanced by pro-hemostatic features (ie, elevated levels of von Willenbrand factor, elevated factor VIII, and low levels of protein C and S). The PT and INR only narrowly reflect one portion of the coagulation cascade, do not reflect a holistic view of the ability to clot, and do not accurately represent bleeding risk.

In fact, at one surgical center, the authors report on a series of 500 consecutive liver transplant procedures performed in the setting of advanced liver disease.³ The mean INR for these patients was 1.8. In spite of this, these authors were able to perform 398 of the 500 surgical procedures without requiring transfusion of any additional blood products and, similarly, without any prophylactic administration of fresh frozen plasma (FFP). In fact, when these authors were designing their surgical protocol for liver transplant, their retrospective case review revealed administration of FFP was actually a risk factor for red blood cell (RBC) transfusion and decreased one-year survival rates. This leads to a second important conclusion regarding the approach to patients with ESLD: FFP should not routinely be used as procedural prophylaxis or treatment for the bleeding cirrhotic with an elevated INR.

Despite running contrary to dogma, this is actually fairly straightforward. If a patient cannot be helped by a particular therapy, they can only be harmed. Therefore, it follows that if patients with ESLD are in a rebalanced state of pro- and anti-hemostatic factors without an innately increased bleeding risk, there remain only the harms associated with FFP transfusion.

Most of the evidence in support of this statement has emerged from case series utilizing clot-formation assays such as thromboelastometry (ROTEM) or thromboelastography (TEG). For example, one pilot study found all but one of 55 patients with elevated INR (≥1.3) in an intensive care unit had, in fact, normal coagulation when tested by ROTEM.⁴ Subsequently, these authors were able to obviate their typical practice of periprocedural transfusion of four units of FFP, and no bleeding complications were observed.

The PT and INR only narrowly reflect one portion of the coagulation cascade, do not reflect a holistic view of the ability to clot, and do not accurately represent bleeding risk.

Specific to cirrhotics, a small randomized trial divided patients into two cohorts: preprocedural transfusion using an INR-based “standard of care” and an alternative TEG-guided pathway.⁵ The mean INR in the study was just under 2, and while only half of the cohort underwent procedures with an elevated bleeding risk, the TEG-guided strategy dramatically reduced FFP and platelet transfusions without any difference in subsequent periprocedural RBC transfusion.

These findings are critically important because a growing body of evidence reflects previously underrecognized risks related to transfusion. Transfusion-associated adverse events cause significant morbidity and mortality while consuming limited resources. Transfusion-associated circulatory overload and transfusion-related lung injury are reported to occur in 1 in 20 to 1 in 50 critically ill patients receiving transfusion and are major causes of transfusion-related fatalities. Transfusion-related immunomodulation also increases the susceptibility to nosocomial infections and sepsis in the critically ill.

Furthermore, in patients with cirrhosis and upper gastrointestinal bleeding (UGIB), aggressive resuscitation is almost certainly harmful. The typical dose required to “correct” an elevated INR of 4 units of FFP adds approximately a liter of high-osmotic intravascular volume. This has the effect of increasing portal venous pressures and can thereby precipitate worsening gastrointestinal hemorrhage. These same harms from transfusion are also seen in RBC transfusions in acute UGIB. In a prospective trial, a restrictive transfusion threshold of 7 g/dL conferred a survival advantage over a more liberal strategy using a threshold of 9 g/dL.⁶ This advantage was maintained in the third of patients with cirrhosis and UGIB and those with variceal bleeds, reducing mortality from 18 percent to 11 percent.

It is clear that our erroneous dependence on the INR for the routine screening and treatment of coagulopathy and bleeding is harming patients and wasting limited resources. Use of the PT and INR should be to answer their narrow spectrum of appropriate clinical questions. We are also compounding the issue by misinterpreting these tests in the setting of ESLD, whose “rebalanced” hemostasis is best reflected only in the newer clot-formation assays. It is time to leave the old dogma behind and look at coagulation through a better lens.

References

1. ACEP DRAFT quality measures for public comment. Measure #3: coagulation studies in adult patients presenting with chest pain with no coagulopathy or bleeding. ACEP website. Accessed Sept. 2, 2016.
2. Tripodi A, Mannucci PM. The coagulopathy of chronic liver disease. N Engl J Med. 2011;365(2):147-156.
3. Massicotte L, Denault AY, Beaulieu D, et al. Transfusion rate for 500 consecutive liver transplantations: experience of one liver transplantation center. Transplantation. 2012;93(12):1276-1281.
4. Durila M, Lukas P, Astraverkhava M, et al. Tracheostomy in intensive care unit patients can be performed without bleeding complications in case of normal thromboelastometry results (EXTEM CT) despite increased PT-INR: a prospective pilot study. BMC Anesthesiol. 2015;15:89. De Pietri L, Bianchini M, Montalti R, et al.
5. Thrombelastography-guided blood product use before invasive procedures in cirrhosis with severe coagulopathy: a randomized, controlled trial. Hepatology. 2016;63(2):566-573.
6. Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368(1):11-21.