Outpatient Oral Anticoagulant Therapy

Initiation and Maintenance Dosing

The administration of a large loading dose of warfarin to initiate therapy (40 to 60 mg was used in the 1960s) is of historical interest only. It induces a rapid, but excessive, reduction in factor VII activity, predisposing patients to hemorrhage in the first few days of therapy, and it fails to achieve a more rapid decline of the other vitamin K-dependent coagulation factors (II, IX, and X). Therapy is properly initiated using an average maintenance dosage (about 5 mg daily) for the first 2 or 3 days. When an immediate effect is required, such as in the treatment of acute venous thrombosis, heparin should be given concurrently with warfarin for at least 5 days. Warfarin treatment should overlap with heparin therapy for a period of 4 to 5 days because it takes that long to lower the levels of prothrombin and factor X, those vitamin K-dependent coagulation factors with longer half-lives and that are mostly responsible for the antithrombotic effect of warfarin. Heparin is usually discontinued when the INR has been in the therapeutic range on two measurements taken at least 24 hours apart. If initiation of treatment is not urgent (e.g., in chronic stable atrial fibrillation [AF]), warfarin can be commenced out of hospital at an anticipated maintenance dosage of about 5 mg daily, which usually achieves a therapeutic anticoagulant effect in about 5 days, although a stable INR may take longer to achieve. Two studies showed that using an initial 10 mg dose of warfarin in outpatients achieved a therapeutic INR more rapidly than using a 5 mg dose without producing a higher rate of excessive anticoagulation. Using a 10 mg starting dose for outpatients is recommended by the 2012 American College of Chest Physicians (ACCP) Evidence-Based Clinical Practice Guidelines. A recent Cochrane Database Systematic Review of the optimal loading dose for warfarin examined 12 studies including studies that used pharmacogenetic dosing and failed to identify an optimal method or dosing schedule between an initial 5 and 10 mg dose. There was evidence in the elderly that using lower initiation doses led to fewer high INRs. The fear of creating a hypercoagulable state in patients with unrecognized protein C deficiency who are not simultaneously given heparin has not been substantiated in patients with AF; however, there is suggestive evidence that in patients with AF started on warfarin without overlapping heparin, there may be an increased risk of early stroke in the first 30 days. Certainly, in patients with a known protein C deficiency or other thrombophilic state, it would be prudent to begin heparin before, or at the same time as, warfarin. A starting dose lower than 5 mg might be appropriate in the elderly, in patients with impaired nutrition or liver disease, and in patients at high risk of bleeding. The physician should be aware of variables that influence the response to anticoagulation in the elderly. The dosage required to maintain a therapeutic range in patients older than 60 years of age has been shown to decrease with increasing age, and older patients are more likely to have other variables that might influence INR stability or the risk of bleeding, such as a greater number of other medical conditions or concurrent drug use. Consequently, it is advisable to monitor treatment more carefully in older patients to maximize their time in the therapeutic range.

Estimation of the maintenance dose is often based on observations of the INR response following administration of a fixed dose of warfarin over an interval of a few days. An individual who rapidly achieves a high therapeutic INR (above 1.5) after two doses of warfarin is likely to require a low maintenance dose. The opposite holds for a patient who shows little elevation of the INR (below 1.5) after two doses. It is now known that a major determinate influencing a patient's response to initial and maintenance dosing is the patient's genotype for the principal enzyme that metabolizes warfarin (CYP 2C9) and the target enzyme through which warfarin mediates its effect (VKORC1). Polymorphisms in the genes for these enzymes account for almost 50% of the variability in dosing requirements. They are also associated with a higher rate of bleeding events. Several recent, large, randomized trials using information about genetic polymorphisms to inform dosing have shown conflicting results. In a North American study, the COAG trial (Clarification of Optimal Anticoagulation through Genetics), investigators randomized 1015 patients in a double-blind fashion to receive initiation of warfarin therapy based on genetic and clinical variables versus initiation of therapy based only on clinical variables. They found no difference in time-in-range between groups in the first 15 days or at 4 weeks (genetic group 45.2% vs. standard group 45.4%; P = .91). Surprisingly, black patients managed by genotype-guided dosing had significantly less time-in-range than nonblack patients managed by standard care in a subgroup analysis. Lastly, they found no significant difference in the rates of the combined outcome of an INR of 4 or more, major bleeding, or thromboembolism based on dosing strategy. At the same time, a smaller European study of genotype-guided warfarin dosing found different outcomes in 455 patients randomized to genotype-guided versus standard dosing where the time-in-range was significantly better for the genotype group (67.4% vs. 60.3%; P < .001). They also found significantly fewer cases of INRs ≥4.0 and the median time to reach a therapeutic INR was 21 days in the genotype-guided group compared with 29 days in the standard dosing group. The divergent outcomes in these two landmark trials may have been due to different patient populations, different genetic frequencies, different dosing algorithms, and other factors. In a systematic review and meta-analysis, Franchini and colleagues primarily assessed the outcomes of major bleeding and thromboembolism in nine randomized trials comparing genotype-guided versus clinically-guided warfarin dosing. They found a significant reduction in major bleeding in the genotype-guided dosing compared with standard dosing (RR, 0.47; 95% CI, 0.23 to 0.96) but no differences in thromboembolism, deaths or the secondary endpoints of time-in-range or patients with an INR ≥4. To further complicate matters, two additional systematic reviews and meta-analyses comparing genotype-guided dosing with non-genotype-guided dosing found differing results. Belley-Cote et al. in an analysis of 12 trials, found no significant differences in the primary outcomes of mortality, thromboembolism, or major bleeding (RR, 0.85; 95% CI, 0.54 to 1.34; P = 0.35). However, they did note a significant difference between groups with time-in-range significantly higher in the genotype-guided group compared with fixed VKA dosing patients, but not when clinical dosing algorithms were used. In an analysis by Stergiopoulos and Brown of nine trials, investigators failed to find significant differences in percentage of time-in-range, patients with an INR greater than 4, or a reduction in major bleeding or thromboembolism. The results of cost-effectiveness studies using pharmacogenetic dosing are of borderline significance and are highly dependent on adverse event variables. There are a number of factors that diminish the value of pharmacogenetics-guided dosing. These include the limited availability of genetic assays, especially with rapid turnaround; cost of genetic assays; complex algorithms needed to utilize genetic information; and conflicting trial results about the usefulness of such information and the overall cost effectiveness. All of this must be compared with the relative inexpense and value of frequent monitoring of a simple INR test. At the present time, pharmacogenetics-based dosing is not recommended by the ACCP guidelines. Until appropriate information is available, frequent monitoring of the INR response is the best means of predicting and establishing the patient's maintenance dose requirements.

INR monitoring is usually performed daily for inpatients (somewhat less frequently for outpatients) until the therapeutic range has been achieved and maintained for at least 2 consecutive days; monitoring is then performed one to two times weekly for 1 to 2 weeks and less often thereafter, depending on the stability of INR results. It should be noted that immediately after hospital discharge is often the most unstable and dangerous time for patients owing to alterations in dietary habits and changes in other medications. Therefore monitoring should be performed frequently for the first 1 to 2 weeks after discharge.

It is well documented that the time-in-therapeutic range (TTR) is highly correlated with efficacy and safety outcomes of warfarin therapy. A number of investigators have identified patient characteristics or other factors that predict instability. These include characteristics such as young age, hospitalizations, pain medications, active cancer ; low weight, and secondary venous thromboembolism (VTE) ; and poor compliance with INR scheduling. Patients who exhibit these characteristics might benefit from more frequent monitoring.

If the PT response remains stable, the frequency of testing can be reduced to intervals as long as every 4 weeks. For patients who have demonstrated stable INR results (no change in dose in 3 months), recent studies suggest that monitoring may be extended to every 8 to 12 weeks. However, there is no guarantee that stable patients remain stable as demonstrated in a report from a large real-world registry of patients on warfarin for stroke prevention in AF. Of 968 patients with a stable INR over a 6-month period (80% or more INRs in range), only 34% remained stable over the subsequent 6 months. Even of those patients with 100% INRs in range over 6 months ( n = 376), only 37% remained stable in the next 6 months. If adjustments to the dose are required, then the cycle of more frequent monitoring is repeated until a stable dose response is again achieved.

Outpatient management of warfarin therapy should aim for simplicity and clarity to avoid patient confusion, poor compliance, and dosing errors that may result in complications. It is recommended that a limited number of warfarin tablet strengths be used in clinical practice, and that patients clearly understand the various dosing patterns that are used, such as alternate-day doses or dosing levels based on days of the week. One must also be aware that several different warfarin products are on the market, which can lead to confusion for the patient. There are generic preparations sold as warfarin, brand preparations sold as Coumadin, and branded generics sold under a different name (e.g., Jantoven). Patients may be given two different preparations and be taking both, not knowing that they are the same drug.

Management of Nontherapeutic International Normalized Ratios

Patients receiving long-term warfarin therapy often have unexpected fluctuations in dose response that require careful management. These may be due to inaccuracy in PT testing, changes in vitamin K intake (increased or decreased vitamin K in the diet), changes in vitamin K or warfarin absorption (GI factors or drug effects), changes in warfarin metabolism (liver disease or drug effects), changes in vitamin K-dependent coagulation factor synthesis or metabolism (liver disease, drug effects, worsening right-sided heart failure with increasing hepatic congestion, other medical conditions), or patient compliance issues (surreptitious self-medication, missed doses, miscommunication about dose adjustment, and so on).

A nontherapeutic (e.g., elevated) INR can be managed by briefly discontinuing warfarin, administering vitamin K, or infusing fresh frozen plasma (FFP) or a factor concentrate. The choice is based largely on the severity of the clinical situation (e.g., degree of elevation of the INR, presence of severe bleeding). Assuming an ongoing normal food intake and reasonable hepatic function, when warfarin administration is interrupted, it takes about 4 to 5 days for the INR to return to the normal range in patients whose INR is between 2.0 and 3.0. The INR will return to normal more quickly in patients requiring a larger daily maintenance dose than in those requiring a lower daily maintenance dose. After treatment with oral vitamin K, the INR declines substantially within 24 hours. Because the absolute daily risk of bleeding is low even when the INR is excessively prolonged, many physicians manage patients with INR values of 4.0 to 9.0 by simply holding warfarin and monitoring more frequently, unless the patient is at a higher risk of bleeding or bleeding has already developed. Vitamin K can be administered by the intravenous, subcutaneous, or oral route. Intravenous injection may be associated with anaphylactic reactions in rare cases, but it does lead to reversal of the INR more quickly than the oral or subcutaneous administration of vitamin K. The response to subcutaneous vitamin K may be unpredictable and sometimes delayed. Several studies confirm earlier reports that oral administration is predictably effective, more so than subcutaneous administration, and has the advantages of safety and convenience over parenteral routes. Ideally, vitamin K should be administered in a dose that will quickly lower the INR into a safe, but not subtherapeutic range, without causing resistance when warfarin is reinstated. High doses of vitamin K, although effective, may lower the INR more than is necessary and lead to warfarin resistance persisting for up to 1 week.

In the presence of significant bleeding, the INR must be reversed immediately. This can be done by replacing the vitamin K-dependent factors using FFP. At least 15 to 30 mL/kg of FFP should be given to have a significant impact on factor replenishment. The amount used should depend on the degree to which the INR is out of range and the severity of the bleeding. In the average 70 kg person, this equates to 4 to 8 units of FFP, which may take up to 6 to 24 hours to order, thaw, and infuse, and in some patients it may put an undue stress on the heart. In addition, the half-lives of the individual vitamin K-dependent factors govern the durability of any response. More FFP might be required. In a patient with life-threatening bleeding, especially intracranial hemorrhage, maximal factor replacement should be achieved in the shortest interval possible. This can be accomplished by using factor concentrates that have high factor concentrations in a small volume. Prothrombin complex concentrates (PCCs), containing the vitamin K-dependent factors, have traditionally been used. Studies have also shown that recombinant factor VIIa (rFVIIa) can reverse the coagulopathy and associated bleeding induced by warfarin, as well as other coagulopathies. Given the expense of these agents, one should use the lowest effective dose to control the bleeding. PCCs are often dosed in a range of 25 to 50 IU/kg, but the specific dose will depend on body weight, degree of INR prolongation, and desired level of correction. The same considerations are relevant for the dosing of rFVIIa, although the range of dosing reported empirically ranges from approximately 20 µg/kg to more than 100 µg/kg. Although these factor concentrates are shown to return the INR to normal more rapidly than with FFP, there is still debate as to how beneficial they are over FFP in terms of outcome. In a randomized, open-label trial comparing 4-factor PCC with FFP in patients taking a VKA and having major bleeding, Sarode et al. were unable to show any significant difference in serious adverse events, death or thromboembolic events between the two treatments. There was an increase in heart failure in the FFP group attributed to the high volume of plasma infusion. PCC infusion resulted in a significantly more rapid correction of the INR as others have shown. Both PCCs and rFVIIa have the potential to induce a prothrombotic state, and thromboembolism has occurred as a result of such therapy. Table 37.4 outlines the 2008 ACCP recommendations for managing patients receiving coumarin anticoagulants who need the INR lowered because of actual or potential bleeding. These recommendations have not significantly changed in the most recent guidelines.

The decision to provide coverage to a patient who comes to one's office with a subtherapeutic INR is similar to a bridging decision. If the thrombogenic risk is so high (such as within the first week or two of treating new DVTs and/or pulmonary embolisms [PEs]) that even a few days at a subtherapeutic level places the patient at high risk, then treating the patient with low-molecular-weight heparin (LMWH) for a few days until a therapeutic INR is achieved with higher doses of warfarin is appropriate. In most cases, this intervention is not needed, and simply raising the dose of warfarin is sufficient to reestablish a therapeutic range in a few days.

Management of Oral Anticoagulation During Invasive Procedures

Clinicians are often confronted with the challenge of managing anticoagulation in individuals requiring noncardiac surgery or other invasive procedures, especially individuals with prosthetic heart valves who are currently taking warfarin (see Chapter 35 ). There is a paucity of critical studies examining the alternative choices for anticoagulation (called bridging therapy ) in this setting, and until recently, there were no robust randomized trials to assess the benefit of bridging therapy versus no bridging therapy. At the present time, based on clinical evidence, there is a strong push against using bridging therapy except in those patients at very high risk of thromboembolism without adequate anticoagulation coverage. When bridging is contemplated, physicians must assess the risk of bleeding from a procedure if anticoagulation is continued versus the risk of thrombosis if anticoagulation is discontinued, as well as the risk of bleeding from an alternative rapidly acting anticoagulant, and the cost of alternative anticoagulation options. Reviews have addressed the management of these patients and the options available, and the ACCP has made formal recommendations. Full-dose intravenous unfractionated heparin (UFH) was originally the agent of choice to bridge a patient during an invasive procedure but has been replaced by LMWH, which offers a less complex alternative because it requires no monitoring and can be given at home. A particular problem with bridging therapy is the risk of post-procedure bleeding if a rapidly acting anticoagulant is started too soon. There is not only a risk of major bleeding but also a risk associated with the delayed reinstitution of long-term therapy because of the bleeding. Investigators have shown a 3-month cumulative risk of major post-procedure bleeding of approximately 3% with bridging versus 1% without such bridging, and selected patient characteristics may be associated with post-procedure bleeding. In a recent study, Douketis et al. randomized 1884 patients with AF at low to moderate risk of thromboembolism off anticoagulants to either LMWH bridging versus no bridging. The outcome of arterial thromboembolism was the same in each group (0.3% vs. 0.4%, respectively; P = .01 for noninferiority). The incidence of major bleeding was 3.2% in the bridging group versus 1.3% in the no bridging group (RR, 0.41; 95% CI, 0.20 to 0.78; P = .005 for superiority). The study population represented a wide spectrum of AF patients at risk except those in the highest risk group and also represented a wide spectrum of invasive procedures. In a recent retrospective analysis of 1178 Kaiser Permanente Colorado patients, mostly with VTE as opposed to AF, investigators found similar findings with a major bleeding incidence of 2.7% in bridged patients versus 0.2% in nonbridged patients (HR, 17.2; 95% CI, 3.9 to 75.1). There was no difference in the rate of VTE.

If bridging therapy is implemented in those at very high risk of thromboembolism, warfarin is usually discontinued 4 or 5 days before the procedure and the INR is allowed to decline. LMWH is started 2 or 3 days before the procedure, usually at a full treatment dose (100 to 150 anti-factor Xa U/kg subcutaneously) once or twice daily depending on the risk of thrombosis, with the last dose omitted the morning or night before the procedure (approximately 12 to 24 hours before the procedure). It is then restarted about 12 hours after the procedure along with warfarin. If the risk of postoperative bleeding from the procedure is high, LMWH can be restarted in 24 to 48 hours, rather than in 12 hours. When the INR becomes therapeutic, LMWH administration is stopped. Table 37.5 summarizes the author's approach integrated with various guideline recommendations for the management of oral anticoagulation during invasive procedures.

Dental procedures represent a particularly common intervention in patients taking anticoagulants. A comprehensive review of the subject indicated that, in most cases, no change in the intensity of anticoagulation is needed (i.e., continue to keep and document an INR of 2.0 to 3.0). Should a need to control local bleeding arise, tranexamic acid or ε-aminocaproic acid mouthwash has been used successfully without interruption of anticoagulant therapy.

Diagnostic Evaluation of Bleeding

When bleeding occurs, especially from the GI or urinary tract, it is important to consider the possibility of a serious underlying occult lesion as the source of bleeding. A number of descriptive studies indicate that the chance of finding such a lesion is approximately 5% to 25%, which supports the need for further investigation in such cases. The significance of hematuria in a patient taking warfarin is less clear. There is a low prevalence of microscopic hematuria in patients not on anticoagulant therapy, and the likelihood of finding an underlying malignancy appears to be less.

Predicting and Managing the Risks of Oral Anticoagulant Therapy

Over the last 20 years, several developments have improved the safety and efficacy of oral anticoagulation. These include defining the optimal intensity of therapy by well-designed randomized trials and standardizing the reporting of PTs using the INR, which have led to more appropriate and standardized therapy. Identification of the risk factors associated with bleeding has also advanced. Time in the therapeutic range is an important measure of the quality of anticoagulation care, and a strong relationship exists between time in the therapeutic range and either bleeding or thrombosis rates as demonstrated by Cannegieter and colleagues, Hylek and coworkers, and others. In addition to time in the therapeutic range, a number of patient characteristics serve as risk factors for bleeding. Knowing these risk factors for an individual patient before initiating anticoagulant therapy or deciding to extend therapy is an essential aspect of care and the balance of risk and benefit. A number of risk indexes have been developed to help the clinician. These include the Bleeding Risk Index of Beyth and colleagues, the HEMORR2HAGES index of Gage and coworkers, the HAS-BLED score of Pisters and colleagues, and the ATRIA score of Fang and coworkers. All of these scoring systems incorporate many common risk factors, including age, history of bleeding, renal insufficiency, anemia, hypertension, and others. Unfortunately, the predictive ability of these scores when looked at by others has not always held up nor is one scoring system particularly superior to another. Moreover, none of them is so specific that other factors can be ignored, and in deciding the risk of bleeding in an individual patient, many factors need to be taken into consideration, including the living environment, other comorbid conditions, patient compliance, and others. Table 37.6 summarizes these bleeding risk indices.