One class of drugs that has been implicated in serious adverse drugs reactions for many decades are the oral anticoagulants, especially warfarin. There are countless cases reports indicating the use of oral anticoagulants is more likely to require hospital admission, and prolonged hospital stays from adverse events such as bleeding (OIG, 2010; Piazza et al., 2011).
The oral anticoagulants have been classified as high alert medications according to the Institute of Safe Medication Practices (ISMP) because they have the potential for harm when used clinically. Many reports have appeared on the risk of bleeding when the anticoagulants are used concurrently with other similar agents (antiplatelet drugs), when the drug treatment is duplicated, in the presence of dosing errors, when there is accidental discontinuation of treatment and when there are problems with monitoring (ISMP, 2016).
Unlike warfarin, the newer oral anticoagulants do not have a long track record in clinical medicine and hence, it is too soon to claim that they are safer. In any case, proactive measures and targeted education should be encouraged to prevent adverse effects associated with the newer anticoagulants. More important, the Joint Commision has designated cautious use of an oral anticoagulant as part of the National Patient Safety Goals (NQF).
How serious is the problem with oral anticoagulants?
From published studies, the incidence of major bleeding in patients treated with warfarin ranges from 0.4% to 7.2% per year. Minor bleeding rates have been found to be high as 15% per year. This wide range in bleeding inside is primarily due to patient-specific comorbid factors. To be consistent in reporting, however, most clinicians define major bleeding as bleeding that required admission, a fatal hemorrhage, bleeding at a critical site like the brain or retroperitoneum or bleeding that required transfusion of at least 2 units of packed red blood cells. A major bleed is associated with a several-fold increase in death for up to 12 months following the incident.
Many studies have shown that the risk of bleeding is increased in patients treated with warfarin. This risk of bleeding in patients treated with atrial fibrillation has been calculated to be about 0.3-0.5% per year and is often associated with intracranial bleeding, which is a major cause of disability and death (Piazza et al, 2011).
Patients with deep vein thrombosis who are treated with warfarin appear to have a higher risk for bleeding than those treated for atrial fibrillation. This is thought to be due to the concomitant comorbid conditions in these patients. The risk of bleeding in venous thrombosis has been calculated to be about 1.31 per 100 person-years, with a case fatality rate of 13.4 %.
Even though intracranial hemorrhage is devastating, it is not the most common site of bleeding following use of oral anticoagulants; the most common site is the gastrointestinal tract. However, the mortality associated with an intracranial bleed is 50%, compared with 5.1% with gastrointestinal tract bleeding (Vaughan et al, 2014)
The association between venous thrombosis and malignancy is well established. Many studies report a high incidence of both minor and major bleeding in patients with cancer who are treated with oral anticoagulants. There appears to be greater fluctuation in the INR in these patients, which could be due to the comorbidity and/or concomitant medications (Abdel-Razeq et al, 2018).
In one large retrospective 5-year study, nearly 48.8% of adverse drug events that involved anticoagulation were due to medication errors. In the same study, the 30-day mortality was increased by 11% of patients who experienced an adverse drug effect due to the anticoagulation drug. Other studies have revealed that emergency hospitalization as a result of bleeding due to warfarin is common. Data on the long-term safety of the newly available oral anticoagulation are very limited or not available. Thus, it is obvious that most of the data on the newer anticoagulants will reflect fewer adverse events or appear positive (FitzHenry et al, 2011; Morais & De Caterina, 2016; Levy et al, 2014).
Prior to 2010, Warfarin was the only approved oral anticoagulant on the market but since then several newer oral anticoagulants have been introduced like dabigatran, apixaban, edoxaban, and rivaroxaban. Warfarin acts by altering the clotting factors 2,7,9,10 and protein C and S. Comparatively, the target-specific oral anticoagulants except for dabigatran work by inhibiting activation of platelets and fibrin clot formation by reversible inhibition of factor Xa. On the other hand, dabigatran works by reversibly inhibiting thrombin which results in decreased thrombin-mediated platelet aggregation (Fenger-Eriksen et al, 2014; Ruff et al, 2014).
What are differences between warfarin and the newer oral anticoagulants?
One of the major difference between warfarin and the target-specific oral anticoagulants is that the latter cannot be monitored using the INR or other blood coagulation tests. In addition, these agents also do not require modification in diet. Cost is also a factor; the newer oral anticoagulants are prohibitively expensive compared to warfarin. So far these newer oral anticoagulants have shown to be safe, but long-term data are lacking. Another important fact is that most healthcare workers who prescribe these oral anticoagulants have little idea on how to deal with them when a patient is booked for surgery, suffers trauma or has a sudden bleeding emergency. Unlike warfarin which can be reversed with plasma or Vitamin K, the factor X inhibitors or direct thrombin inhibitors have reversal agents that are more expensive and less readily available (Fenger-Eriksen et al., 2014; Ansell et al., 2013).
Another difference between warfarin and the newer oral anticoagulants is the time to reach peak therapeutic effect. Warfarin often takes 4-6 days to reach the peak effects, but the newer oral anticoagulants have peak therapeutic effect in less than 24 hours. However, one area of similarity between the two class of drugs is that they both have the potential for significant drug interactions.
Finally the after dosing with warfarin, monitoring can also be affected by the diet. Foods that contain vitamin K like kale and spinach can delay the peak therapeutic effect. On the other hand, the target-specific oral anticoagulants are highly protein bound and this makes it difficult to remove them even with dialysis.
Where do most adverse events regarding anticoagulants occur?
Which oral anticoagulants are used today?
The newer oral anticoagulants, in general, have been associated with lower rates of major bleeding and fewer intracranial hemorrhagic events. There is also one study which shows that the use of dabigatran in elderly patients and those with kidney dysfunction may increase the risk of intracranial bleeding.
What factors are usually considered when selecting an oral anticoagulant?
What types of medication errors have been associated with oral anticoagulants leading to adverse events?
It should be noted that warfarin has been the mainstay of anticoagulant therapy for decades, while the other oral anticoagulants have been introduced over the past few years. Therefore, in most of the published literature, the number of adverse events involving warfarin are much greater than the number of events involving newer anticoagulants (Michaels et al., 2010).
When and Why does bleeding occur after use of oral anticoagulants?
Studies reveal that the first 90 days are the most variable because the INR is still labile and this increases the risk of bleeding. The same dose of warfarin can have unpredictable effects on the INR in different individuals making it difficult to develop a standardized dosage. Further, many patients require alternating doses, and this makes it difficult to have a standard dose. Because the risk of bleeding is highest within the first three months, there should be more frequent monitoring of the INR during this period to prevent bleeding episodes (Gomes et al., 2013).
Another risk factor is the intensity of anticoagulation. When the INR is more than 2, the risk of bleeding is much greater than when the INR is less than 2. However, when the INR is less than 2, this also increases the risk of stroke in patients with atrial fibrillation. Since warfarin cannot be stopped in atrial fibrillation and an INR less than 2 is not effective to prevent a stroke, other strategies need to be employed like discontinuing platelet agents like aspirin if there is a significant concern for bleeding.
Patient characteristics can also increase bleeding. Studies show that the bleeding risk is increased with the following:
Further patients admitted with sepsis and other hypermetabolic state tend to have an unpredictable response to warfarin, and the bleeding potential is always present.
Drugs that can increase the risk of bleeding include use of:
Over the years several enzyme genetic variations that affect the metabolism of warfarin have been identified. Polymorphism of the cytochrome P4502CP enzymes slows down the metabolism of warfarin, which increases the risk of bleeding. Similarly, alterations of the vitamin K epoxide reductase complex subunit one gene and the VKORC1 enzyme also alter metabolism. The clinical significance of these genetic variations is known because there are limited data. Plus not all laboratories are established to detect these genetic changes, thus making its universal use unrealistic (Daneshjou et al., 2013).
So how do we reduce the risk of adverse events involving oral anticoagulants?
Almost every healthcare institution has faced issues with the use of oral anticoagulation, and over the years many have developed strategies to reduce the risk of harm to patients. While education of healthcare workers is important, it also relies on the fact that individual will remain compliant with the education on prevention of harm. Most experts feel that strategies that are based on the entire system (i.e., involve everyone from the pharmacist to the eventual patient) are far more effective as they function at many levels and the errors can hopefully be picked up before the drug even gets to the patient. Some of the risk prevention strategies designed to prevent harm from oral anticoagulants include the following:
Bleeding Risk Assessment Tools
Even though oral anticoagulants are effective drugs, there is concern that these drugs may be cause bleeding problems while on therapy. Hence, today several bleeding risk assessment tools based on patient risk factors have been developed. These tools are now being incorporated into clinical decision making. These tools help the clinician better quantify the individual risk of hemorrhage and help classify the patient as low or high risk. The one thing to appreciate is that all the presently available bleeding risk assessment tools were developed for patients treated with warfarin and to date, it is not known whether these same tools can be applied to the newer oral anticoagulants (Shoeb et al., 2013).
The three widely used Bleeding risk Assessment tools include the following:
All of these bleeding risk tools are relatively easy to use than their predecessors. With all these tools, one can obtain the common risk factors from the patient medical history or routinely obtained laboratory investigations that are performed in patients with atrial fibrillation or deep vein thrombosis. The Atria and HAS-Bled score utilize fewer variables and no genetic testing is necessary. The HAS-Bled tool has been validated in several clinical trials and is currently the most frequently used tool to screen for risk of bleeding. Further, the Has-Bled tool also incorporates labile INR, which is obtained following the start of warfarin therapy.
The atria bleeding risk tool is best suited for use in high-risk patients. The one major limitation of the Atria tool is that the score does not include comorbid conditions like blood pressure or use of medications that can increase the risk of bleeding. Currently, Atria is not very widely used and is awaiting clinical validation.
The Hemorr2hages score is comprehensive, and while it takes into account the comorbidity and medications that can increase bleeding, it also requires genetic testing for liver enzymes. The downside is that genetic testing is not always practical in emergency situations and many hospitals lack the capability to conduct such testing.
Bleeding Risk Tools in Clinical Practice
So far there is no evidence that one risk bleeding tool is better than the other and the choice to use one is based on experience and personal preference. Many of the risk factors in the different tools overlap. Older age is a common risk factor with all risk bleeding tools. Other similar risk factors in the tools include the history of bleeding, renal disease, and anemia. All of them require the need to monitor INR and place control limits on the levels. However, these risk tools may not be applicable to patients on the newer oral anticoagulants who do not require INR monitoring. In addition, genetic testing is not only prohibitively expensive; it may not be available in most healthcare institutions. While several validation studies have been performed in subgroups of participants in clinical trials, the overall observation is that these bleeding risk tools only have a modest predictive value in patients with atrial fibrillation.
To enhance the effectiveness of these bleeding risk tools, it is important to understand that even if these tools predict bleeding, will anticoagulation therapy be discontinued? Remember the negative consequence of thromboembolism are also serious and in many cases outweigh the consequence of hemorrhage. Today no risk tool has been designed to predict the risk of hemorrhage in the brain. Although some risk tools may predict the risk of an intracranial bleed, the threshold at which this bleeding will occur is not known. In reality, these risk bleeding tools are more useful in patients at low risk of thromboembolism where the net benefit of anticoagulation is small, and the risk of bleeding is real. These bleeding risk tools may also identify low-risk patients who can be told that their risk of bleeding is low
When there is an error in anticoagulation dosing, the result can be either hemorrhage or thrombosis. Hemorrhage is most common in the Gastrointestinal tract and can be profuse. Often the site is a prior peptic ulcer, angiodysplasia or AV malformation. The Thrombosis usually results in an ischemic stroke, mesenteric ischemia or an acutely cold leg. These diagnoses are made from imaging studies, and histopathology is not required to confirm the diagnosis.
If the dose of the anticoagulation is high, the patient may complain of internal bleeding after a few days. Obvious signs of bleeding are not present in the first 24 hours after ingestion of the drug. Besides intracranial hemorrhage, the patient may have:
In all cases of anticoagulation associated adverse events, the following tests are needed:
Other studies depend on the patient presentation.
If the patient is bleeding, the patient may require immediate discontinuation of the oral anticoagulants. Recombinant factor Vlla has been used to lower the INR. The most readily available source of this is plasma transfusion. Vitamin K can be used to reverse warfarin-induced bleeding. Fresh frozen plasma is often administered to patients with bleeding due to the newer anticoagulants, but is of limited efficacy. Some institutions give prothrombin complex concentrate in patients with severe gastrointestinal bleeding, this can be beneficial in severe, life-threatining bleeds related to vitamin K antagonists. There is no evidence to suggest newer oral anti-coagulant bleeds respond to prothrombin complex, concentrate.
If bleeding is severe, intravenous resuscitation is necessary. Neurological signs will need to be monitored to look for the presence of an intracranial bleed. Epistaxis may require nasal packing.
The differential diagnosis includes:
The prognosis of patients who bleeding following use of oral anticoagulants depends on the comorbidity, age, and level of INR. Fatalities are not uncommon if the patient develops an intracranial bleed. If the bleeding is minor from the mucous membranes, hematuria, epistaxis or ecchymoses, recovery is common with few complications. However, if the bleeding is severe, it can result in catastrophic gastrointestinal or intracranial hemorrhage. Others may develop hemorrhage in the pleural, pericardial and abdominal cavity.
For those who develop prosthetic valve thrombosis, the prognosis is guarded. If thrombolytic therapy fails to dissolve the thrombus, emergent valve replacement surgery is required- which also carries a significant morbidity.
Complications of oral anticoagulants include:
Education and Information
Oral anticoagulants are classified as high-risk medications, and if errors are made in dosing, monitoring or inappropriate administration, there is a real risk of bleeding, which appears to be greater with the use of warfarin compared to the newer oral anticoagulants. When the dose of the oral anticoagulants is excessive, bleeding is known to occur which adds to morbidity and cost of care. But if the dose is not therapeutic, then there is a risk of thrombosis which carries a far greater morbidity and mortality than hemorrhage. For those patients who do require oral anticoagulation, one may use bleeding risk scores to determine the risk of bleeding
So far the limited data indicate that the newer oral anticoagulants are safer compared to warfarin, but long-term data on their safety are lacking. Hospitals should establish precautions at multiple levels to help minimize the occurrence of adverse events with these agents.
Strategies to Lower risk of bleeding
Monitoring of Adverse Drug Events