C. Gregory Elliott, M.D., of the Pulmonary Division at LDS Hospital in Salt Lake City, Utah, discussed the prevention of venous thromboembolism (VTE) in surgical patients. There is a high prevalence of VTE among hospitalized patients with potentially fatal outcomes. The efficacy of prophylactic methods in reducing fatal VTE cannot be studied due to the low frequency of events, and there are few trials evaluating symptomatic VTE. The majority of the data for prevention of VTE examines asymptomatic VTE, since the frequency allows for adequately powered trials (Figure 1).

In 1998, an alert was issued by the FDA concerning the use of LMWH with epidural or spinal anesthesia or spinal puncture. The alert was in response to a small number of patients who developed epidural or spinal hematomas when LMWH was used concurrently with regional anesthesia. Therefore, when considering the most appropriate prophylactic modality to use, consider the Fifth Annual ACCP Conference guidelines on the use of LMWH with epidural or spinal anesthesia. With postoperative LMWH, it is advised to delay LMWH therapy for two hours after the removal of the catheter. Any decision to use LMWH with an indwelling epidural or spinal catheter should be made with extreme caution and avoided if possible.

In summary, the Fifth ACCP Consensus Conference guidelines for prophylaxis of VTE following THR are based on level 1 evidence. The guidelines recommend postoperative LMWH with an alternative agent being preoperative warfarin to an INR of 2 to 3. Prophylaxis should be continued for at least 7 to 10 days, and longer prophylaxis may be considered in patients with continued risk, such as those with prolonged immobility, those who are elderly, and those who are obese.

The recommended prophylaxis in patients sustaining major trauma at a high risk of bleeding is the use of calf-thigh IPC. In individuals with a low risk of bleeding, enoxaparin, 30 mg twice-daily, is recommended.

More evidence is needed concerning the prevention of VTE after elective neurosurgery. Unlike orthopedic surgery, elective neurosurgery is a heterogeneous group of procedures and the specific procedure needs to be considered in the decision of how to initiate prophylaxis. There are certain situations that place patients at higher risk of VTE, including intracranial surgery, intracranial surgery for malignancy, a long duration of surgery, and patients with residual immobility of the legs or paresis.

General surgery encompasses the most heterogeneous group of procedures and is not easily standardized. Table 5 lists the aggregate prevalence of DVT detected by mandatory venography in general surgery. The Fifth ACCP Consensus Conference came up with a scheme to categorize general surgery patients according to their risk of VTE (Table 6). The risk spectrum for VTE from lowest to highest risk is listed in Table 7, along with the range of recommended prophylaxis measures.

Dr. Russell Hull, director of the Thrombosis Research Unit in Calgary, Alberta, Canada, discussed the treatment of established VTE. The current evidence is clear that intravenous (IV) heparin is obsolete for the treatment of VTE due to the introduction of LMWH. LMWH is a small molecular weight compound that is made enzymatically or chemically from heparin. Because LMWH compounds are made differently and vary in their anti-Xa-to-IIa ratio, they are not interchangeable unless proven so clinically. Only enoxaparin and tinzaparin have been compared in a prophylactic trial and found to be equivalent despite different anti-Xa-to-IIa ratios.

Some problems with heparin include the need to be dosed adequately and to be monitored using the activated partial thromboplastin time (APTT) to ensure its effectiveness. The dose of heparin is determined by the attainment of a therapeutic APTT value; therefore, frequent dose adjustments may be needed. Inappropriate dosing of heparin, as reflected by an APTT of less than 1.5, is associated with an increased risk of recurrent VTE of 25% compared to a risk of 2% with an APTT greater than 1.5. To help ensure proper therapeutic APTT values, most institutions have implemented a heparin dosing protocol. Alternatively, the anticoagulant effect of LMWH has been correlated to body weight and, therefore, is dosed based on the patient's weight and does not require monitoring. Using anti-Xa levels to monitor the anticoagulant effect of LMWH is not necessary since there is no correlation between these levels and clinical outcome. Monitoring anti-Xa levels has been suggested when LMWH is used in obese patients, since dosage is determined by total body weight, and in renal failure patients, due to possible accumulation. However, since the anti-Xa levels have not been correlated to clinical outcomes, it may be better to use heparin in these patients until further information is available for LMWH dosing in obesity and renal failure.

LMWH also has a higher bioavailability as compared to heparin. Depending on the dosage, heparin may only have a bioavailability of 30% due to binding to plasma proteins and endothelial cells, thus leading to unpredictable levels. LMWH also has a longer half-life for anti-Xa activity compared to heparin that allows for less frequent dosing. Heparin can be given subcutaneously (SC) similar to LMWH; however, it must be given at least twice-daily and monitoring is required. Comparative trials are underway to evaluate weight-adjusted SC heparin and LMWH. The recent approval of LMWH for the treatment of VTE in the U.S. was aided by its ease in administration and pharmacokinetic advantages.

Enoxaparin is the LMWH currently approved in the United States for the treatment of VTE. The recommended dosages are 1 mg/kg every 12 hours or 1.5 mg/kg once-daily (inpatient use only). The dose of enoxaparin given daily is unique compared to the other LMWH compounds. Of the other LMWH compounds that were historically studied twice-daily and that have gone to once-daily dosing, the once-daily dose is the twice-daily dose doubled. Initially, investigators felt once-daily dosing for enoxaparin would not be ideal since the anti-Xa levels seemed to be attenuated by some 8 to 24 hours. Due to this belief, twice-daily dosing was studied in the majority of the trials. Reducing the total daily dose of enoxaparin to the once-daily regimen is an incompletely studied situation.

In the U.S., the cost savings associated with home treatment of DVT with LMWH will predominantly result from the reduction in hospital days. It has been found that the cost of diagnosis of a VTE is not dependent on the length of hospitalization. However, in terms of treatment, a hospital stay of less than 24 hours in the U.S. will result in significant cost savings when compared to a hospital stay beyond 24 hours. Even if a patient has to be admitted initially, discharging them quickly to home treatment will still result in a reduction in costs. Therefore, the goal is to avoid or abbreviate hospitalization and implement home treatment.

Some items to consider with the use of LMWH home treatment are whether to use standard of care based on guidelines and evidence-based medicine or on information from the package insert. Other considerations include basing decisions on literature support of treatment approaches, having physician involvement in the patient selection, having patient agreement to receive home treatment, and documenting delegation of who will be following the patient. There are a variety of outpatient models for DVT treatment programs. The models that will be covered are those that have been described in the literature. Some types of models include a thromboembolic service, an anticoagulation clinic, ambulatory care clinic or day care area, emergency department or urgent care area, fast-track areas, or a physician office program. Many of these models will incorporate home self-injection and a community follow-up program. However, some models may have patients report daily to receive their injections and monitoring. Sites should optimize their clinical resources and clinical expertise to develop a site-specific model for their institution.

The core personnel for an outpatient DVT treatment model should include medical staff, pharmacy staff, home nursing liaisons (in some models), home nursing representatives, and laboratory support staff. Informal reports indicate anywhere from 50% to 90% of patients will elect to self-inject while others would prefer injection via a family member, a friend, or daily visits to a healthcare provider. Before choosing a model, a cost analysis should be done to take into account staffing and space considerations. Funding and reimbursement systems should be researched (that will often dictate the model that will be used) and a treatment algorithm should be developed.

There are many reasons to consider using an anticoagulation clinic model. After the results of the home treatment trials were available, many anticoagulation clinics were developed to incorporate the use of LMWH at home. The use of an anticoagulation clinic is logical based on the fact that warfarin initiation is crucial for the duration of LMWH therapy. Anticoagulation clinics have mechanisms for direct physician referrals, experience with comprehensive case management, availability in tertiary and non-tertiary care centers, team approaches, and experience with initiation and long-term management of warfarin. An anticoagulation clinic can also incorporate patients who would like to self-inject, as well as those who need to come in daily for injections.

Based on the types of patients evaluated in the trials, those who would be good candidates for home treatment for DVT are patients with uncomplicated, documented DVT and who are capable of self-injections or are eligible for nursing support. As more evidence becomes available, the selection criteria may expand. Examples of patients who should not receive LMWH therapy at home include patients with PE who have hemodynamic instability, active bleeding, high risk of bleeding complications, history of a prior gastrointestinal bleed, recent surgery or major trauma, severe renal failure, a hypercoagulable state, potential for nonadherence, significant comorbidities, pregnancy, body weight over 150 kg, and heparin-induced thrombocytopenia.

In determining which LMWH to use, consideration should be given to the evidence supporting the use of the LMWH for treatment of VTE. At this time, the best evidence is available for enoxaparin, nadroparin, and tinzaparin; the data on dalteparin are less strong. Currently in the U.S., only two LMWH products are available, enoxaparin and dalteparin, with only enoxaparin being FDA-approved for the treatment of DVT. In choosing a LMWH, the institution's formulary and cost considerations should be taken into account; as more LMWH compounds become available, this will become more of an issue. The availability of syringe sizes for the products should also be considered.

Pharmacists can be involved in the selection of eligible patients, as well as be part of the patient education process. There are booklets and videotapes available. Key components of patient education should include information on the disease state and the intent of therapy, including whether it will involve short-term issues associated with LMWH or long-term issues associated with warfarin. Patients need to understand the importance of self-monitoring, especially if they are to self-inject medication, and the importance of follow-up. Patients should also receive education on the drug supply and adequacy of the supply upon discharge. Where pharmacists can also play an integral role is in the comprehensive case management of the patients, including working together with the patient's family, physicians, and home care agencies. Other areas that pharmacists can participate in include documentation, quality assurance for all parts of the outpatient therapy, and analysis of clinical outcomes and costs. Clinical outcomes to evaluate include bleeding rates, thromboembolic events, and cost analyses that consider hospital days avoided, complications avoided, medication costs, and personnel costs. All facilities should develop drug-specific instructional objectives that may be assisted by company-produced and institution-edited objectives. Reading level and language should be considered, so that the format is appropriate for the patient population.

Reimbursement issues obviously impact the home treatment process. Unfortunately, there is no single answer for reimbursement for home treatment, and much will depend on the institution and the patient-payer mix. It is crucial to determine the patient's medical and drug benefit coverage and any managed care limitations that may be imposed on home treatment. Different policies may have limits for home nursing visits and patient homebound status. As part of the JCAHO requirements, patients need to be informed of their financial responsibilities.

David Rosenbloom, Pharm.D., Director of Pharmaceutical Services in Hamilton, Ontario, Canada, discussed opportunities for clinical pharmacists to transfer research evidence into practice. A barrier to using research evidence in clinical practice is understanding the research. Providing insight into clinical trials, clinical epidemiology and biostatistics, pharmacoepidemiology, and pharmacoeconomics is needed to enhance understanding. Other barriers include knowing the difference between efficacy and effectiveness and knowing how to determine if a study applies to your patient population. Efficacy refers to whether a drug or intervention works in a clinical trial or a restricted population. Effectiveness refers to whether a drug or intervention works within the general population. Determining the relevance of a study's results can usually be accomplished by examining the study's inclusion and exclusion criteria. If the inclusion and exclusion criteria are not followed when applying the results of a trial to the population, then the data are not evidence-based.

In trying to change physician or pharmacist prescribing behavior, educating them via the use of practice guidelines and opinions of leaders in the subject area has a small effect. A greater effect is achieved through feedback obtained from clinical experience. Feedback is characterized as providing an opinion on what would be best to do or what may have worked better. Feedback is most useful when analyzed prospectively. It is more successful if a prescribing pattern can be influenced in real time rather than in looking backward and analyzing what should have been done. Involving physicians in the process of changing prescribing behavior and setting new standards helps to achieve the desired goals. It is not possible to just impose a change on physicians. Physicians will respond to authority to a degree. Often financial incentives will make a difference in changing behavior, if that luxury is available. Administrative interventions can be used, such as restricting a medication to a specialty and/or placing a barrier or limit on its use, altering order forms to provide physician reminders, and legislating change. Interventions to induce change often can work if the physician and the pharmacist work together.

A number of primary care physicians were polled as to their reasons for not providing the standard of care that is accepted as level 1 data (see Table 1). Sixty-seven percent felt it was due to a lack of an effective system to remind the patients, and 44.5% felt is was due to a failure to remind the physician. The solution may be as simple as implementing a system to automatically remind the physician.

In many cases the ideal requisite for the action phase is the use of care management tools. They provide the opportunity to implement changes utilizing the measurement of outcomes and costs. They also provide a good background to practice clinical research and can be used for educational purposes. However, there may be potential problems. Practitioners may not be interested in change and may be imbedded in the precontemplative state. In addition, a needs assessment may not have been undertaken to match the needs of the institution; practitioners may be in compliance with the institution's ideal practices and, therefore, not need to seek change; and the education program may be too long for the intended audience. One needs to explain the benefits and risks and look at the requirements for each stage of the change and plan for the appropriate people and resources. One-on-one detailing will also work, but this is not the most efficient way to achieve change. Combining the readiness for change model with the care management tools is the ideal way to implement a change.

In conclusion, one should work with both the leadership and practitioners to achieve an institutional approval of the process. A performance monitoring and management system should be developed to know who is complying and then to make it inevitable for all to comply. The change should be integrated across a continuum of care so the same principles apply when going from hospital to home. An efficient infrastructure and an educational process are also essential. Knowledge of critical appraisal is important, and one needs to read and understand clinical trials. One needs to understand how to gain buy-in by using the evidence-based approach and the readiness to change concepts and to combine all the tools to achieve the appropriate patient outcomes.

Greg Samsa, Ph.D., of the Duke Center for Clinical Health Policy Research in Durham, North Carolina, discussed the impact of a systematic approach for oral anticoagulant management and the MAST study. The models of care of anticoagulation management include physician office-based services, anticoagulation services (ACS), and patient self-management (PSM). A major advantage for the physician office-based service is that there are fewer people involved. Physician office-based anticoagulation management can be strengthened by computer-based reminder systems, computer-based tracking systems, and point-of-care testing devices. The idea behind ACS is improved training and consistency of application. The tasks are delegated to providers who are experts in the field with improvement in the quality of care. With PSM, the major advantage is that the patient can monitor therapy more frequently. Given these three models, no single model of care will be best for every patient and provider, and one needs to think of the best model for his or her environment and patient population.

Definitive research (level 1 trials) evaluating the various models of care is lacking due to the difficulty in designing such trials. It is difficult to evaluate outcomes such as event rates since the rates are usually low and it may not be feasible to design such a large study. Therefore, intermediate outcomes are often evaluated instead, such as time-in-range. When examining data retrospectively, assessing whether the provider has made a good decision is difficult since the rationale is usually not sufficiently documented. To look prospectively, a system of care would need to be in place to avoid haphazard care and achieve high-quality care. Therefore, to design a trial, the experimental and control groups should be changed to high-quality models of care.

Another unpublished trial randomized patients with idiopathic VTE to treatment with oral anticoagulants for three months or one year. At one year, the three-month treatment group had a recurrence rate of 7.5% compared to 1.5% for the one-year treatment group. However, after reaching the two-year mark, the risk of recurrent events equalized. Treatment for one year prevented six events per 100 patients. The annual rate of major bleeding was about 3% for extended therapy. The current recommendation is to treat patients with a first VTE due to a transient risk factor for three to six months and to treat patients with an idiopathic VTE for 6 to 12 months.

Management of patients with antiphospholipid antibodies (APA), or specifically anticardiolipin antibodies, is a very complicated area that is compounded by the lack of a standardized methodology for measuring the antibodies. There are many different types of APA and not all of them may be relevant to increasing the risk of recurrent VTE.

In patients with VTE, the risk of major hemorrhage with warfarin therapy (INR 2 to 3) during the first three months is about 3%, with a yearly rate of 2% to 5% based on an ongoing trial of extended therapy. It is important to note that risk of bleeding will vary depending on the indication. A decision analysis of the data on VTE found that the efficacy of continuing warfarin varies with time from the index event and with patient characteristics. More reliable estimates of the baseline risk and the experienced quality of life (QOL) are needed. The QOL in patients with VTE will depend on factors such as incidence of adverse events and the perceived burden of treatment or impairment on daily activities. In patients with atrial fibrillation, it has been found that long-term warfarin does not adversely impact QOL.

There are various factors that influence the decision on how long a patient should take oral anticoagulant therapy for VTE (Table 10). An appropriate practice will incorporate research evidence, good clinical expertise, and patient preference. One does not want to say to patients that they are on warfarin for life, but rather that they are on it indefinitely. Long-term anticoagulation should be re-evaluated at regular intervals, at least yearly, due to the need to incorporate changes in the available evidence and to reassess the risk of major bleeding or other features that may make the risk-benefit assessment differ.

There are ongoing clinical trials in North America that will help clarify the risk-benefit of oral anticoagulation. In Oklahoma, a specialized center from the National Heart, Lung and Blood Institute is evaluating extending warfarin treatment in a first episode of VTE for three years versus six months. By the end of 2000, the study will have at least 100 patients with the Leiden mutation. The PREVENT study is evaluating patients with idiopathic VTE randomized to an extended duration of treatment to a less intense INR (1.5 to 2.0). The SOFAST study is randomizing patients (with a DVT due to a transient risk factor) to four weeks or three months of treatment. The ELATE study is evaluating patients with an idiopathic VTE treated for three months and then randomized to extended treatment either with an INR of 2.0 to 3.0 or 1.5 to 2.0.

Dr. Lou Fiore of Boston VA Medical Center in Boston, Massachusetts, discussed the primary and secondary prevention of myocardial infarction (MI) and the roles of warfarin and aspirin. Epidemiologic data have found that the concentration of factor VII and fibrinogen in the blood predicts ischemic heart disease development in five years as strongly as cholesterol levels. With plaque rupture, a thrombus forms that may become totally occlusive and may result in an MI. Alternatively, the lesion may become non-occlusive and be lysed, with the plaque returning to its original size, or it may not be lysed, with the thrombus becoming incorporated into the plaque structure and, thus, resulting in progression of atherosclerosis. Atherosclerosis does not form linearly, but with staggered steps based on thrombus formation and organization. Fibrinogen and factor VII may be associated with atherosclerosis; if there is a higher procoagulant load, larger thrombi may form with plaque rupture and with sporadic growth of the plaque. Therefore, if factor VII is lowered and platelets are inhibited, this process may happen less intensely or less frequently, slowing the process of atherosclerosis.

Prior trials that evaluated the combination of aspirin and warfarin used high doses of aspirin and found large increases in bleeding, particularly gastrointestinal. At the present time it is known that 80 mg/day of aspirin causes less gastrointestinal bleeding, suggesting that a therapeutic effect may be seen with less bleeding. The amount of warfarin will then impact the risk of bleeding when given with aspirin. Prior trials have found that an INR of 2.8 to 4.8 prevents death and reinfarction post-MI and that there is probably a dose-response relationship.

There are currently ongoing combination therapy trials evaluating higher INR ranges. The WARIS-2 is evaluating warfarin, aspirin, or the combination with an INR of 2 to 3 (results due in August 2000). ASPECT-2 is evaluating warfarin, aspirin, or the combination with an INR of 2.5 to 3.5 (results expected in 2001). These studies should clearly answer if there is any role for combination warfarin and aspirin post-MI.

Dawn Havrda, Pharm.D., BCPS, from the University of Oklahoma College of Pharmacy in Oklahoma City, discussed the prevention of systemic embolism in patients with left ventricular dysfunction. Twenty to fifty percent of patients with LV dysfunction die of a sudden cardiac death (SCD). The etiology of SCD is usually attributed to a fatal arrhythmia; however, some of these deaths may be due to a fatal systemic embolism affecting the cerebrovascular or pulmonary circulation. From the existing evidence, the incidence of systemic embolism in LV dysfunction is between 0.9 and 42.4 events per 100 patient-years (pt-yrs). This wide variation is due to the heterogeneity of the patient populations studied, and the weighted average is about 1.9 events per 100 pt-yrs.

Virchow's triad (abnormalities in blood flow, the vessel wall, and blood constituents) can explain the mechanism behind how a thrombus forms in LV dysfunction. With LV dysfunction, a low cardiac output, aberrant blood flow through dilated chambers, and reduced contractility result in abnormalities in blood flow and stasis of blood with pooling of coagulation factors. Studies have found that patients with LV dysfunction have abnormalities in the vessel wall manifested as a defective endothelium that can lead to platelet adhesion and subsequent thrombus formation. Patients with LV dysfunction have been shown to be hypercoagulable compared to normal controls, and the degree of hypercoagulability is related to the severity of heart failure. Patients have increased activation of the coagulation system, higher levels of activated factor X, and higher amounts of thrombin generation and plasmin formation. Studies have found a poor association between echocardiographic visualization of a thrombi and subsequent risk of embolization. The embolic events are likely due to small thrombi located in variable locations within the heart that are not easily visualized. Therefore, the echocardiogram has little use in predicting the risk of a systemic embolism in LV dysfunction.

Anticoagulation has been investigated in LV dysfunction since 1948. Anticoagulation has been found to reverse the markers of hypercoagulable studies, but it is not known if this reversal results in a reduction in clinical events. To date, there are no large prospective randomized controlled trials investigating the use of anticoagulation in LV dysfunction. The problems with the existing data are the use of small sample sizes and small event rates, the retrospective nature, and inconsistent results.

The second group of trials evaluated patients with nonischemic and ischemic causes of LV dysfunction. The incidence of systemic embolism was 0.9 to 4.3 events per 100 pt-yrs without anticoagulation therapy. Anticoagulation therapy was not effective in reducing the risk of systemic embolism. However, the trials were not randomized, so high-risk patients were more likely to receive anticoagulation therapy. There was also no control over the intensity of the anticoagulation therapy. The studies were short in duration with the maximum follow-up being two years. The studies found that the stroke rate did not depend on the functional status or etiology of LV dysfunction. Since some of the trials found a low incidence of systemic events, the question remains whether anticoagulation therapy should be used, since the risks may outweigh the small benefit.

Subgroup analyses from the heart failure trials did not find a reduction in systemic embolism with the use of anticoagulation therapy. Some problems with this evidence are that it is from subgroup analyses and the trials were not designed to evaluate the efficacy and safety of anticoagulation therapy. Due to a lack of randomization within these trials, there was no control over patient characteristics. High-risk patients tended to receive anticoagulation therapy. In addition, there was no control over the intensity of therapy. Bleeding events were also not recorded, and the classification of outcomes was not optimal and may have led to under- or over-reporting of systemic embolic events.

There are certain risk factors that will increase the chance of a systemic embolic event and may be useful in making treatment decisions. The two risk factors identified have been atrial fibrillation (AF) and a low ejection fraction. Five AF trials have identified a history of a recent episode of heart failure or echocardiographic evidence of LV dysfunction as independent predictors of stroke. Alternatively, the LV dysfunction studies are not consistent with a relationship between a higher number of systemic embolic events with AF. However, since many of the trials considered AF to be high risk, patients were placed on anticoagulation therapy. This action makes it difficult to find an association between AF and LV dysfunction and an increased risk of systemic embolism.

Many of the trials found an inverse relationship between the severity of ejection fraction and the risk of systemic embolism. The SAVE trial found that for every 5% reduction in ejection fraction, the risk of stroke increased by 18%. In addition, there was an increased chance of stroke in patients with ejection fractions less than 28%. The SOLVD trial found that for every 10% reduction in ejection fraction, the risk of SCD increased by 47%. The problem with using ejection fraction as a risk factor for systemic embolism is the difficulty in defining low ejection fraction. The combined studies have found patients with varying severity of LV dysfunction to be at increased risk.

To evaluate whether or not to use anticoagulation therapy in individuals without any of these risk factors, the pros and cons of treatment need to be considered. Clinical trials have indicated that the incidence of systemic embolism is 1.9 events per 100 pt-yrs without anticoagulation therapy. Unfortunately, these trials have not provided consistent evidence on the benefits associated with anticoagulation therapy. From the AF trials, an INR of 2.0 to 3.0 is associated with an annual incidence of major bleeding of 1.3% and intracranial bleeding of 0.3%. Obviously the rates would be expected to increase as the INR exceeds 3.0. A potential problem in LV dysfunction is hepatic congestion that may be associated with heart failure. It has been hypothesized that hepatic congestion could lead to alterations in warfarin metabolism and fluctuations in the INR, making the patients more prone to over-anticoagulation and bleeding events. Due to lack of prospective data, it is not known if this is a realistic problem, but hopefully ongoing studies will help to define this risk. Overall, the risk of bleeding approximates the risk of systemic embolism, making it difficult to determine if the benefits outweigh the risks.

Stuart Haines, Pharm.D., BCPS, CDE, from the University of Maryland School of Pharmacy in Baltimore, Maryland, discussed advances with newer oral anticoagulants. There are some disadvantages to the current antithrombotic agents that would warrant looking for alternatives. The available agents produce an unpredictable patient response and, therefore, require a system to monitor these patients. The system is expensive to run and unavailable to the majority of patients. This system also relies on a set of tests (e.g., APTT, PT/INR) that are not totally accurate. There is also the disadvantage of the potential for hemorrhagic complications, as well as drug-drug interactions, drug-food interactions, and high anxiety about substitution of products with the use of warfarin. Payment of services for this system is not in place and remains a problem for many practitioners.

The ideal anticoagulant agent would be highly effective and comparable to the current agents. It would have a low incidence of hemorrhagic complications and side effects. An ideal agent would have long-term safety and be safe in special populations, such as patients with renal or hepatic dysfunction, or patients who are pregnant, elderly, or children. It should also be orally administered and produce a predictable antithrombotic response to reduce or eliminate the need for therapeutic monitoring. If monitoring is needed, then it should be accurate and reliable with point-of-care testing available. Drug interactions should be minimal. Any agent should be tested for its efficacy and safety when given in combination with other antithrombotic agents. An ideal agent would be inexpensive or cost-effective.

The mechanism of action of oral heparin is the same as the intravenous form of heparin. The bioavailability of heparin is poor and unpredictable via the subcutaneous route. Oral absorption is not appreciable due to its large size and polarity, but may be improved if combined with a vehicle that allows its absorption, such as a vehicle referred to as SNAC (N-[8(-2hydroxybenzoyl) amino] caprylate). SNAC facilitates the passive transcellular transport of large molecules, such as heparin. In animal models, SNAC plus heparin has shown some potential. SNAC does not cause damage to the endothelial lining of the stomach, it prolongs the APTT, and it prevents VTE in a rat model. SNAC or oral heparin alone was not found to prolong the APTT in animals, but the combination produced a large increase in the APTT. In a model of VTE in rats, SNAC alone did not prevent VTE, but oral heparin alone decreased the rate to 44% even though the APTT was not prolonged and the combination reduced the rate of VTE even further.

One published, randomized controlled trial (phase 1) evaluated oral heparin in humans. It was a dose escalation study that evaluated the effect on the APTT, anti-Xa activity, tissue factor pathway inhibitor, and tolerability. The trial evaluated SNAC in combination with 30,000, 60,000, 90,000, 120,000, and 150,000 units of heparin. To prolong the APTT to 1.5-times control, doses of 90,000 to 150,000 units of heparin plus SNAC were required. The variability in the response was not reported. SNAC was not well tolerated in the trial, with the most common side effects being nausea and emesis. Therefore, due to the poor tolerability of SNAC, a better tolerated vehicle may be more desirable. LMWH with a vehicle such as SNAC will probably be evaluated in the future.

Sulodexide is a glycoaminoglycan similar to heparin. It is made up of LMWH (80%) and dermatan sulfate (20%). Its mechanism of action is via antithrombin and heparin cofactor II. Sulodexide is available in Europe as an oral agent and as an injectable preparation that is usually given intramuscularly (IM). Some of the pharmacokinetic characteristics include a long half-life (11 to 26 hours), daily or twice-daily dosing, and renal (55%) and biliary (23%) elimination. The oral absorption of sulodexide is ill defined; when radio-labeled oral sulodexide is given, it appears to be 100% bioavailable but has different pharmacodynamic effects than the IM route. The IM route prolongs the APTT and has anti-Xa activity, but oral administration does neither. Perhaps when sulodexide is given orally the heparin particles are degraded so that it is not able to prolong the APTT, but the dermatan sulfate component remains active.

Other properties of sulodexide include inhibition of platelet activation, antihyperlipidemic activity with a reduction in triglycerides, and antiproliferative activity. It also reduces fibrinogen and plasminogen activator inhibitor levels, increases clot lysis time, and reduces blood and plasma viscosity. It is used in Europe for peripheral vascular occlusive disease. It also reduces event rates for cardiovascular disease, with one study showing substantial improvement in recurrent coronary events with sulodexide IM for 1 month and orally for 11 months. There is also evidence in Europe that it is effective for cerebrovascular disease, and it has been studied in the prevention and treatment of VTE with comparable effects to heparin. It has also been examined in diabetic patients and has some benefit in reducing their risk of coronary events.

Over the past 20 years there has been efforts to produce compounds with direct activity against factor IIa (thrombin), such as hirudin, or against factor Xa. Unfortunately, with the exception of the hirudin compounds, many of these agents have not been successful. Factor IIa and Xa are serine proteases that are a diverse class of enzymes found universally within individuals. Any compound that is produced must be specific for just IIa or Xa and not for the whole class of serine proteases. Initially, covalently bonding serine protease inhibitors that would irreversibly bind to these enzymes at their active site were investigated. It was very difficult to produce such an inhibitor that is specific for just IIa or Xa. These agents were also vulnerable to rapid metabolism, which led to short half-lives. These compounds also had slow binding kinetics, so that it took time for them to attach to their binding site and often they were eliminated before they attached. Therefore, the approach was changed to develop non-covalent bound Xa and IIa inhibitors.

The TAME (N-(-tosylarginine methylester) derivatives, such as argatroban, were the first to be developed to inhibit factor IIa. The problem was that argatroban had poor oral absorption, and other companies have investigated similar compounds with better bioavailability but with short half-lives. Another compound in this class, inogatran, has demonstrated poor oral bioavailability.

A second class, the peptidominietics, is comprised of very small compounds that are highly selective for factor IIa and have demonstrated higher levels of oral absorption. One compound (L-375,378) has been found to have good oral bioavailability in dogs (90%) and monkeys (60%), with half-lives of three hours in dogs and one hour in monkeys. No data are currently available about its success in humans.

Another approach is to try to specifically inhibit factor Xa. Daiichi, a Japanese company, is investigating a compound (DX-9065a) that has shown good bioavailability in the rat model (> 50%) and an effect similar to warfarin's in preventing VTE in rat models. Due to its specific inhibition of Xa, it also appears to cause less blood loss. In the baboon model, the bioavailability was not as good as with the rat model, but the compound has proceeded to human trials. Another compound (YM-60828) by Yamanouchi is similar to DX-9065a, with good activity in the squirrel monkey, prolongation of the PT and APTT, and a relatively good half-life and bioavailability (20%). However, it is chemically unstable and needs to be stored by special procedures. A salt of this compound was found to prevent VTE as well as warfarin in a rat model and also produced less blood loss than warfarin. It has a wide therapeutic range, so it could be given in large or small doses, and it demonstrated no observable drug interactions in a rat model.

In conclusion, the promise of oral heparin is limited by the variation in anticoagulant response and the intolerability of the vehicle. Oral LMWH may be more promising due to its superior pharmacokinetics over heparin. Sulodexide has a wide breadth of indications, but it is still under investigation. The direct thrombin and Xa inhibitors are also currently undergoing investigation and questions remain about their use.