Description:
Patients who are prescribed chronic warfarin anticoagulation need ongoing monitoring that has generally taken place in a physician's office or anticoagulation clinic. Home prothrombin monitoring with a U.S. Food and Drug Administration (FDA)-approved device is proposed as an alternative to office or laboratory based testing.

Warfarin is an effective anticoagulant for the treatment and prevention of venous and arterial thrombosis. Chronic warfarin therapy is recommended in all patients with mechanical heart valves and in some patients with chronic atrial fibrillation (i.e., patients with one high-risk factor or more than one moderate risk factor). Patients with mechanical heart valves are frequently prescribed anticoagulants at higher levels than patients given anticoagulants for other indications, which puts them at higher risk of complications from warfarin therapy. Appropriate levels of warfarin anticoagulation are monitored with periodic prothrombin time measurements, as measured by the International Normalized Ratio (INR). For example, an INR result greater than 3 indicates a higher risk of serious hemorrhage, while an INR of 6 indicates an increased risk of developing a serious bleed nearly seven times that of someone with an INR less than 3. In contrast, an INR less than 2 is associated with an increased risk of stroke. Therefore, monitoring of the prothrombin time is recommended to ensure that the prescribed dosing regimens result in INRs within the therapeutic range.

There are at least three sites/methods of monitoring anticoagulation:

• Physician's office (80 percent) — usually once a month
• Anticoagulation clinics (20 percent) — usually once every 2 to 3 weeks
• Home prothrombin time monitors (<5 percent)

In order for home prothrombin time monitoring to be effective, patients need to be appropriately trained and be able to generate INR test results comparable to laboratory measures. Moreover, the clinical impact of home prothrombin time monitoring is related to improved warfarin management. Specifically, home prothrombin time monitoring permits more frequent monitoring and self-management of warfarin therapy with the ultimate goal of 1) increasing the time that the anticoagulation is within a therapeutic INR range (intermediate health outcome); and 2) decreasing the incidence of thromboembolic or hemorrhagic events (final health outcome). Home self-monitoring is typically associated with some form of self-management of warfarin therapy. In some cases, the patient may be supplied with treatment algorithms and instructed to alter the dose based on the results of self-monitoring. In other cases, the patient may be instructed to telephone the results of the self-monitoring and receive further telephonic instructions on warfarin dosage.

Regulatory Status:
In January 2007, the CoaguChek XS^® system (patient self-testing) was cleared for marketing by the FDA through the 510(k) process. The FDA determined that this device was substantially equivalent to existing devices, including the CoaguChek SX system (professional). Other than a labeling change, the device is identical to the professional version of the CoaguChek XS system. The patient self-testing system is intended for self-monitoring of prothrombin time in patients who are on a stable regimen of anticoagulation medications. Other examples of FDA approved devices are ProTime^®, Avocet^® and Rubicon^®.

Policy:
At-home monitoring of chronic warfarin therapy may be considered MEDICALLY NECESSARY in patients who require continuous anticoagulation for chronic conditions. These conditions include, but are not limited to, the following:

•  Patients with mechanical heart valves 
•  Chronic atrial fibrillation

Before initiation of at-home monitoring, patients must have undergone anticoagulation management for at least three months.

Medical Appropriateness:

• Patient must be HOMEBOUND due to medical reasons for the duration of treatment; OR
• Must not be within 30 miles of a hospital laboratory, free-standing laboratory or physician's office that performs INR testing and with the potential for non-compliance; AND
• Meet all of the following:

1. Duration of warfarin treatment is expected for six months or longer.
2. Prothrombin time monitoring is required at least once a week.
3. Documentation that the individual or caregiver has been trained and is capable of performing the test accurately.
4. Documentation of action plan for abnormal test results.

Home prothrombin time monitoring for anticoagulation management of other conditions/diseases is considered INVESTIGATIONAL.

Any device utilized for this procedure must have FDA approval specific to the indication, otherwise it is considered INVESTIGATIONAL.

Rationale:
The preferred study design to evaluate home prothrombin time monitoring is a randomized trial comparing home prothrombin monitoring to either monitoring in a physician's office or monitoring in specialized coagulation clinics. As with any monitoring technology, one would ideally want to isolate the contribution of the monitored data itself from the possible impact of increase patient education or contact with health professionals that is typically associated with more intense monitoring. Final health outcomes would preferably focus on the incidence of hemorrhagic or embolic events. However, due to the low incidence of events, published studies often report instead the intermediate outcome of time spent in the therapeutic range of warfarin, and measured by the International Normalized Ratio (INR). Prior research has established a strong link between time outside the therapeutic range and adverse events such as bleeding and thromboembolism. Therefore, time in the therapeutic range is an intermediate outcome for true health outcomes.

This policy was originally created in 1997 and was updated regularly with searches of the MEDLINE database. The most recent literature search was performed for the period January 2011 through January 2012. Following is a summary of the literature date:

Home monitoring during maintenance therapy
The majority of studies of home PT monitoring include patients who are already being treated with Coumadin, and therefore evaluate home monitoring in the setting of maintenance therapy. A 2010 Cochrane review evaluated the impact of self-monitoring and self-management, which additionally involved having patients interpret the results and adjust the dose of medication themselves. The authors identified 18 randomized controlled trials (RCTs) (total of n=4,723 participants). The review identifies trials on self-monitoring, 11 trials on self-management, and one trial that reported both outcomes. (In reporting the findings, the "home monitoring" will refer to self-monitoring or self-management). Three trials included only patients with mechanical heart valves (MHVs), two trials included patients with atrial fibrillation (AF), and 13 trials included patients with a mix of indications for anticoagulation. A pooled analysis of data from all 18 trials found that, compared to standard therapy, home monitoring reduced thromboembolic events by half (relative risk [RR]: 0.50, 95 percent confidence interval [CI]: 0.36 to 0.69). This difference was statistically significant, p<001. The intervention effect was larger in the self-management studies (pooled RR; 0.47, 95 percent CI: 0.314 to 0.70, p=00003) than in the self-monitoring studies (pooled RR: 0.57, 95 percent CI: 0.32 to 1.00, p=0.05). A pooled analysis of data from nine trials found a statistically significant reduction in mortality with home monitoring compared to standard therapy (pooled RR: 0.64, 95 percent CI: 0.46 to 0.89, p=0.007). In terms of potential harms, a pooled analysis of data from 14 trials did not find a significant difference in the risk of major hemorrhage with standard therapy compared to home monitoring. The RR was 0.87, 95 percent CI: 0.66 to 1.16, p=0.34. Compared to standard therapy, however, there was a significantly reduced risk of minor hemorrhage with home monitoring. Pooling data from 14 trials, the RR was 0.64, 95 percent CI: 0.54 to 0.77, p<0.00001. Although this study reported that self-monitoring and self-management resulted in better outcomes without increasing the risk of major hemorrhage, home monitoring was not feasible for up to half of patients requiring anticoagulant therapy due to patient refusal, doctor recommendation, or inability to complete training.

A similar meta-analysis was published in 2011 by Bloomfield and colleagues. (2) The newer meta-analysis included a large RCT by Matcher and colleagues (2010) (described in more detail below) which was not available at the time the Cochrane systematic review was conducted. Study inclusion criteria included RCTs with adult subjects that compared home monitoring to monitoring in a physician’s office or anticoagulation clinic; studies included adults receiving long-term (>3 months) therapy. The systematic reviews identified 22 trials; five on self-monitoring only and 14 that included self-management. Eight studies were limited to patients with mechanical heart valves and the other 14 included patients with various indications. Three studies enrolled inception cohorts (patients who had started oral anticoagulation therapy in the previous three months), eight studies did not enroll inception cohorts, and eight studies were mixed or unclear about this inclusion criterion. (See later section on home self-monitoring from the beginning of treatment.) In a pooled analysis, there were significantly fewer major thromboembolic events in the self-monitoring and self-management group (99 of 4,004 patients, 2.5 percent) compared to the standard treatment group (149 of 3,755, 4.0 percent), OR: 0.58 (95 percent CI: 0.45 to 0.75). Rates of major bleeding events did not differ significantly in the two groups. There were 283/4,061 (7.0 percent) events in the home-monitoring group and 300/3,806 (7.9 percent) in the standard treatment group; OR: 0.89, 95 percent CI: 0.75 to 1.05. Similar to the Cochrane review, the authors noted the low rate of study participation in patients who met preliminary eligibility criteria. The authors did not conduct separate analysis of studies that did and did not enroll inception cohorts.

A 2012 meta-analysis by Heneghan and colleagues used individual patient data; otherwise, the design was similar to the other published meta-analyses. (3) The investigators searched for RCTs comparing self-monitoring or self-management of oral anticoagulation by adults compared to management by a physician or anticoagulation clinic. This review did not discuss the issue of whether or not home monitoring occurred in the initial three months of anticoagulation therapy. The meta-analysis identified 21 eligible trials, but the authors were not able to obtain adequate data on 10 of the trials, therefore, they included data on 6,417 participants from the other 11 trials. In a pooled analysis, there was a statistically significant reduction in thromboembolic events in the home prothrombin time (PT monitoring group compared to the standard therapy group [hazard ratio (HR): 0.51, 95 percent CI: 0.31 to 0.85]. There was not a significant difference between groups in the rate of major hemorrhagic events (HR: 0.88, 95 percent CI: 0.74 to 1.06) or death (HR: 0.82, 95 percent CI: 0.62 to 1.09).

Representative RCTs enrolling patients initially managed in a physician’s office or anticoagulation clinic are described below.

In 2010, Matchar and colleagues published findings from a large non-blinded multicenter Veterans Administration-sponsored randomized trial. (4,5) The trial, called the Home INR Study (THINRS), included patients who were taking warfarin because of MHVs and/or AF and were expected to be on warfarin indefinitely (operationally defined as 2 years). To be eligible, patients needed to participate in home monitoring training and pass a competency evaluation. A total of 3,643 were trained and 2,922 (80 percent) were randomly assigned to self-testing with an approved device once a week (n=1,465) or monthly clinic-based testing (n=1,457). There were 237 (8 percent) patients who required caregiver support to perform INR testing. Clinics in the study were required to be "high-quality" testing sites, defined as having a designated staff member for patient evaluation and follow-up, using a standard local protocol, and performing INR testing about once a month. Patients in the home monitoring group contacted the clinic with their test results. All participants had quarterly in-person evaluations. About 98 percent of the patients were men, 92 percent were white, and 92 percent had received anticoagulation treatment for at least three months. There were 1,201 (82 percent) patients with AF and 351 (24 percent) with MHVs; 1,113 (76 percent) had AF without MHV. Loss to follow-up was low and similar in the two groups, about 1 percent.

The primary efficacy outcome was time to the first major event (stroke, major bleeding episode, or death) and used Kaplan-Meier analysis; there were 8,730 patient-years of follow-up. A total of 271 (19 percent) of patients in the self-testing group and 285 (20 percent) in the clinic-testing group experienced a major event. There was no statistically significant difference in time to first major event between groups; the unadjusted hazard ratio (HR) was 0.88 (95 percent CI: 0.75 to 1.04). Moreover, there were no significant between-group differences for the individual components of the primary outcome. For example, the rate of death was 152 (10 percent) in the self-testing group and 157 (11 percent) in the clinic-testing group, unadjusted HR: 0.91, 95 percent CI: 0.73 to 1.12. However, the time during which the INR was in the therapeutic range, a secondary outcome, was somewhat higher in the self-testing group than the clinic-testing group (absolute difference, 3.8 percent, 95 percent CI: 2.7–5.0 percent); this difference was statistically significant, p<0.001. The investigators had hypothesized that home testing would be superior to clinic-based testing. They interpreted their finding of no difference between groups in the primary outcomes as an indication that there is no substantial negative effect of self-testing, and they recommended self-testing for patients who have limited access to high-quality anticoagulation clinics.

In a study published in 2005, Fitzmaurice and colleagues randomized 617 patients older than age 18 years and receiving warfarin (approximately 50 percent for AF) to intervention or routine care. (6) Patients receiving intervention used a point-of-care device to measure INR twice a week and a simple dosing chart to interpret their dose of warfarin. No significant differences were found in percentage of time in the therapeutic range between self-management and routine care (70 percent vs. 68 percent). Self-managed patients with poor control before the study showed an improvement in control that was not seen in the routine care group. Nine patients (2.8/100 patient years) had serious adverse events in the self-managed group, compared with seven patients (2.7/100 patient years) in the routine care arm. The authors concluded that, with appropriate training, self-management was safe and reliable for a sizeable proportion of patients receiving oral anticoagulation and may improve the time spent in the therapeutic range for patients with initially poor control.

In a European study published in 2005, Menendez-Jandula and colleagues reported on 737 patients with indications (approximately 50 percent had AF) for anticoagulant treatment. (7) The self-management group (n=368) received simple instructions for using a portable coagulometer weekly and self-adjusting treatment dose. The conventional management group (n=369) received usual care in an anticoagulation clinic (monthly measurement and control of INR, managed by hematologists). The median follow-up period was 11.8 months. The unadjusted percentages of in-range INRs were 58.6 percent in the self-management group and 55.6 percent in the conventional management group (95 percent CI for difference, 0.4 to 5.4 percentage points). Twenty-seven patients (7.3 percent) in the conventional management group and eight (2.2 percent) in the self-management group had major complications related to anticoagulant treatment; the unadjusted risk difference for major complications between groups was 5.1 percentage points (95 percent CI: 1.7 to 8.5 percentage points). This trial was performed at only one center and was not blinded. The dropout rate in the intervention group was 21 percent.

Conclusions: Home INR monitoring is feasible for some patients on warfarin, but a large proportion are unable or unwilling to perform home monitoring. For patients who are able to self-monitor and who are suitably trained, evidence from multiple RCTs demonstrates that home monitoring is associated with better INR control. The impact on outcomes such as thromboembolic events is less certain. Some RCTs report large reductions in thromboembolic events, but these studies are limited by high rates of dropouts due to inability or unwillingness to self-monitor. In other studies, such as the THINRS trial where compliance was established prior to randomization, there were no differences in thromboembolic events. These results suggest that the differences in thromboembolic events may be explained partially or totally by patient differences, since patients who are able to self-monitor are likely to differ in important ways from patients who are not able to self-monitor.

Home self-monitoring from the start of outpatient treatment
Initiation of anticoagulation in the outpatient setting is more challenging than maintenance therapy. During the first few months of anticoagulation, INR levels require frequent monitoring and adjustment of Coumadin dose. The risk for adverse events is also higher during this period, particularly for hemorrhagic complications. As a result, the outcome of hemorrhagic complications is more important during this period compared with maintenance therapy.

Fewer studies report health outcomes in patients being home PT monitored at or near the time they initiate oral anticoagulation therapy. As noted above, three of 21 studies included in the Bloomfield et al. meta-analysis (2) focused on patients who had started oral anticoagulation therapy in the past three months. One of these studies, Beyth et al., was not able to adequately isolate the impact of home PT-monitoring on health outcomes, since it included a multicomponent intervention of which home monitoring was only one part. In addition, only 46 of 163 (28 percent) patients randomized to the intervention group were able to monitor their PT themselves. (8) For another 50 (31 percent) patients, a spouse, other relative, or visiting nurse monitored their PT at home, while 36 (22 percent) were monitored conventionally. The other two RCTs that included "inception cohorts" are briefly described below:

Findings from the first 600 patients who completed the two-year follow-up of the Early Self-Controlled Anticoagulation Trial (ESCAT) from Germany were published in 2001. (9) All patients had undergone mechanical heart valve replacement. A total of 295 patients were in the conventional group and 305 were in the INR self-management group. Self-management included training in use of the CoaguChek device six to 11 days after surgery and received a device after successful training. In the conventionally managed group, 62 percent of recorded INR values were within the stipulated range (2.5 to 4.5) during the entire observation period compared to 79 percent of INR values in the self-management group, p<0.01. Differences in the rate of complications did not differ significantly between groups. Rates of hemorrhagic events were 2.6 percent in the conventionally managed group and 1.7 percent in the self-management group, and rates of thromboembolic events were 2.1 percent and 1.2 percent, respectively (p>0.05).

In 2009, Hamad and colleagues in the Netherlands published an RCT with 62 patients who had elective mechanical aortic valve replacement. (10) Patients were randomly assigned to conventional management or self-management using the CoaguChek device. Self-monitoring training was provided three weeks after surgery. After training, patients were supervised by clinic staff, and they had to pass an exam on self-monitoring before using the CoaguCheck device at home. A total of 58/62 (94 percent) completed the one-year follow-up, 29 in each group. The mean number of INR values per patient within the target range (2.5-4.5) was significantly higher in the self-management group (72.9 +/- 11 percent) than in the conventionally managed group (53.9 +/- 14 percent), p=0.01. In addition, the mean number of days the INR value for each patient was outside the target range was significantly higher in the conventionally managed group (28.6 +/- 14) than the self-monitoring group (22.2 +/- 10), p<0.001. There were no significant differences between groups in postoperative complications or the mortality rate within 1 year of the operation. 

In addition, Gardiner and colleagues published a prospective cohort study from the U.K. in 2009 evaluating the acceptability and efficacy of home prothrombin time (PT) monitoring from the start of treatment. (11) A total of 188 of 318 (59 percent) consecutive patients referred for oral anticoagulation were considered eligible for self-monitoring and were offered the choice of self-monitoring or routine hospital anticoagulation. Of these, 84 (44 percent) chose to self-monitor. Seventy-two patients completed training and started self-monitoring; 26 completed their course of oral anticoagulation during the study period, and 42 were still self-monitoring at the end of the study period; four patients had incomplete data. Data from 67 out of the 84 (80 percent) patients who chose to self-monitor were included in the analysis. Among the 104 patients who elected for routine anticoagulation, only 88 (85 percent) had sufficient data available to be included in the analysis. The median percent time spent in therapeutic range was 71 percent (95 percent CI: 64–75%) for the 67 self-monitoring patients and 60 percent (95 percent CI: 55–63 percent) for the 88 routine care patients; this difference was statistically significant (p=0.003). In the self-monitoring group, the incidence of adverse events included 1.7 major bleeds per 100 patient-years of follow-up, 8.4 minor bleeds per 100 patient-years, and 3.4 thromboses per 100 patient-years. In the routine care group, there were 5.4 major bleeds per 100 patient-years of follow-up, 16.2 minor bleeds per 100 patient-years, and 1.4 thromboses per 100 patient-years.

Conclusions: There is some RCT evidence on home monitoring in the initial treatment period, but the quantity of evidence is less and some of these trials also have high dropout and noncompliance rates. The evidence supports that time in therapeutic INR is improved with home monitoring, but the evidence on adverse events is not sufficient. In particular, the studies have low numbers of hemorrhagic complications and may be underpowered to detect clinically important differences in that outcome.

Summary:
Data from multiple RCTs consistently demonstrate that the use of self-monitoring in patients who were initially managed in a clinical setting results in an increased time in the therapeutic range. It is likely that time in therapeutic INR is associated with improved outcomes based on prior research. Some studies also report a lower rate of hemorrhagic or embolic events, but this evidence is limited by high dropout and noncompliance rates that may have created imbalances between treatment groups. The evidence includes monitoring in several chronic conditions such as mechanical heart valves, chronic atrial fibrillation, and deep venous thrombosis, and therefore comparable results should be able to be obtained in other similar, but less prevalent, conditions that require continuous anticoagulation. Thus, based on the evidence and clinical context, home prothrombin time monitoring may be considered MEDICALLY NECESSARY for patients with chronic conditions that require continuous oral anticoagulation with warfarin who are able to self-monitor and who have undergone initial clinic-based anticoagulation management for at least three months.

The evidence is insufficient to conclude that the use of initial self-monitoring improves the net health outcome. RCTs with sufficient large samples that report health outcomes are needed to more thoroughly evaluate the safety and efficacy of home prothrombin time monitoring from the start of treatment. Therefore, the use of home monitoring during the initial treatment period is considered INVESTIGATIONAL.

Practice Guidelines and Position Statements:
In 2008, the American College of Chest Physicians (ACCP) published evidence-based guidelines on management of vitamin K antagonists. (12) These guidelines included a statement that patient self-testing can be an effective alternative treatment model in patients who are suitably selected and trained.

Medicare National Coverage:
On December 20, 2007, CMS issued a decision memo on home prothrombin time monitoring. (13) The memo expanded the indication for home prothrombin time-monitoring from only including patients with mechanical heart valves to also include patients with chronic atrial fibrillation and deep venous thrombosis. The 2007 memo specified that coverage was for beneficiaries who had been taking anticoagulants for at least three months before use of the home device, had undergone an educational program on anticoagulation management, and were conducting self-testing with the device no more frequently than once a week.

References:

1. Connock M, Stevens C, Fry-Smith A et al. Clinical effectiveness and cost effectiveness of different models of managing long-term oral anticoagulation therapy: a systematic review and economic modeling. Health Technology Assessment 2007; 11: number 38.
2. Proposed Decision Memo for Prothrombin Time (INR) Monitor for home Anticoagulation Management (CAG-00087R).
3. Fitzmaurice DA, Murray ET, McCahon D et al. Self-management of oral anticoagulation: randomized trial. BMJ 2005; 331(7524):1057.
4. Menendez-Jandula B, Souto JC, Oliver A et al. Comparing self-management of oral anticoagulant therapy with clinical management: a randomized trial. Ann Intern Med 2005; 142(1):1-10.
5. Koertke H, Minami K, Bairaktaris A et al. INR self-management following mechanical heart valve replacement. J Thromb Thrombolysis 2000; 9 (suppl1):S41-5.
6. Gardiner C, Longair I, Pescott MA et al. Self-monitoring of oral anticoagulation: does it work outside trial conditions? J Clin Pathol 2009; 62(2):168-71.
7. Ansell J, Hirsh J, Hylek E et al. Pharmacology and management of the vitamin K antagonists: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest 2008; 133(6 Suppl):160S-198S.
8. CMS Decision Memorandum: Home prothrombin time (INR) monitor for anticoagulation management.
9. Proposed Decision Memo for Prothrombin Time (INR) Monitor for Home Anticoagulation Management (CAG-00087R).

^® is not a registered mark of this health plan.