Patient Outcomes Template:



Oral Anticoagulation To Prevent Stroke

in Patients with Atrial Fibrillation

Pathologist’s Summary

Atrial Fibrillation

Atrial fibrillation (AF) is disorganized electrical activity of the muscle wall of the upper chambers of the heart. The disorganization of electrical impulse causes loss of coordinated atrial contraction, which, in turn, allows mural thrombi to form along the inner surface of the atrial wall. Loss of the organized atrial signal also renders the heart’s ventricular rhythm “irregularly irregular” as the temporal variation between atrial signals to the ventricular pacemaker becomes random. In this situation, pieces of left atrial thrombus may tear off, be propelled through the left ventricle, enter the systemic circulation, and reach the cerebral vasculature. There such a thromboembolus can occlude an arterial vessel, cut off blood supply to the downstream brain tissue, and precipitate an ischemic stoke. AF is, in fact, a potent risk factor for ischemic stroke. In the United States today, about 80% of strokes are ischemic, and, in 2000, AF was the major risk in 15% of all strokes. AF is divided into valvular and nonvalvular types. Nonvalvular AF, the predominant form in the U.S., is the type found in patients with anatomically normal mitral heart valves. Non-valvular AF is, by itself, an independent risk factor for stroke, increasing the latter condition’s annual incidence 6-fold. Nonvalvular AF’s prevalence has also been increasing dramatically as the American population ages: it appears in only 0.5% of people < 60 years old but is found in 16% of men and 12% of women > 75 years of age. In this discussion, when AF is mentioned, nonvalvular AF is the type described.

Oral Anticoagulation

During the 1990’s, six randomized, controlled trials of warfarin anticoagulation demonstrated that oral anticoagulation can reduce the annual risk for ischemic stroke in AF patients three fold, from about 4.5 to 1.5%. These dramatic findings have not led to corresponding implementation of oral anticoagulation for AF patients in clinical practice. The reluctance to implement is linked to fear of hemorrhage as a complication, despite a well-documented low annual hemorrhage complication rates in study patients taking warfarin for this indication. These rates average 1.3% and range from 0.15% in patients 60 years old or less to 1.6% among patients greater than 85 years old. For major non-cerebral hemorrhage and for intercranial hemorrhage, the bleeding rates in studies are not different among treated patients from those measured in the randomized trials’ control groups. The reluctance to treat with warfarin also appears despite the superior performance of warfarin compared to aspirin anticoagulation in AF patients, in head-to-head placebo controlled trials. In one trial warfarin decreased the risk of stroke twice as much as aspirin did, without a significant difference in complication rates. Nevertheless, a large, more than 13,400 patient, prospective study of an HMO-based adult population, with a mean age of 72 years and a slight (57%) male predominance, quantified the underuse of oral anticoagulation in AF patients. Most of the AF patients in the study cohort [a group sharing common features, followed over time] had other risk factors in addition to AF that also increased their likelihood of stroke: elevated blood pressure (51%) congestive heart failure (31%) coronary artery disease (30%), diabetes mellitus (17%) and previous ischemic stroke (9%). Conversely, only 17% of AF patients in this ambulatory population had potential contraindications to warfarin therapy, these included: previous gastrointestinal or intracranial hemorrhage (5%), and previous fall requiring medical attention (4%). Nevertheless, in this population, only 55% of AF patients with no contraindications were prescribed warfarin; even in the population subgroup of “ideal candidates” for warfarin treatment those with the best benefit: risk ratio, only 62% received oral anticoagulation.

Stroke

Stroke itself is the third most frequent cause of overall mortality in the United States. It accounted for about 150,000 deaths/year in the mid 1990’s. More strikingly, it is the leading cause of adult disability: of the more than 550,000 new strokes each year, only 20-27% are fatal: In 1992 more than two million stroke survivors resided in extended care facilities in the United States. There they received rehabilitation and care at a total estimated cost (in 1993 dollars) of 30 billion; 17 billion of this amount were direct costs for medical care. In this context, effective stroke prevention in AF patients has a major impact on the overall outcomes of mortality, morbidity, and cost.

Four Obstacles to Oral Anticoagulation in AF Patients

Four factors contribute to the underuse of warfarin in AF patients. The first factor is a heuristic: [having to do with how people discover and learn about a topic]. The prevention of stroke, the major benefit of warfarin treatment in AF, is a negative finding, an achievement of which it is hard to be sure, outside the setting of a controlled study; hemorrhage, the major complication of warfarin treatment, is, in contrast, a positive finding easily attributed to the drug intervention in an individual case, even through the latter can be shown to be a no more than a background risk to a population in placebo controlled studies. This sort of ascertainment bias [inclination that inhibits impartiality in the consideration of relevant cases] makes the benefit: risk ratio of warfarin treatment in AF appear worse, in unsystematic personal experience, than it is demonstrated to be, in randomized, controlled clinical trials.

The second factor contributing to warfarin’s underuse is pharmacologic [having to do with the absorption metabolism, excretion, actions and interactions of the drug]: the drug effect of warfarin is increased or decreased by a variety of common medications or diet constituents: acetominophen is the most frequently incriminated medication that increases warfarin’s anticoagulant effect, while leafy green vegetables and avacados are common diet constituents that have frequently been shown to decrease the drug’s effect.

The third factor that obstructs warfarin use is secular [pertaining to the effects of time or aging on a state of affairs]: as mentioned above, AF most frequently presents in patients older than 75 years . These patients are thought to be more variably sensitive to warfarin’s anticoagulant effect than are younger patients. Whether this is true or not, getting an older patient to a stable warfarin regimen seems more difficult and prolonged a process, than it does in younger patients.

The last obstructive factor is practical [regarding the actual logistics of delivering medical care]: warfarin treatment must be monitored through the entire course of therapy. Moreover, it must be monitored using an indirect index of drug effect, the prothrombin time (PT), rather than a direct measure of drug presence, like a warfarin drug level.

The Laboratory’s Role in the Monitoring of Oral Anticoagulation

The need for monitoring places the clinical laboratory in the midst of AF patients’ warfarin therapy for the prevention of stroke. In the last two decades, the prothrombin time’s (PT’s) test reagent, thromboplastin [an extract of (usually) rabbit brain and/or a mixture of rabbit brain and lung tissue] has been standardized. Standardization has been achieved by measuring the thromboplastin reagents’ tendency to trigger clotting, that is, its ‘sensitivity’ to the presence of the cascade of clotting proteins in patient plasma samples. This measure is quantified in the International Sensitivity Index (ISI). A thromboplastin with a low ISI is comparatively labile to clotting, while a reagent with a high ISI is relatively resistant. If this reagent variation is not taken into account, then the same degree of warfarin effect in the patient will appear different in the test, not as a consequence of drug effect or patient response, but only because of different thromboplastins. The variation among thromboplastins is folded into the estimation of anticoagulant effect by a translation (“normalization”) of the patient:control prothrombin time ratio. This ratio is normalized by raising the quotient of the ratio to the power of the employed thromboplastin’s ISI. This product is called an International Normalized Ratio (INR), a ratio without units that has become the usual format for reporting prothrombin time as a measurement of anticoagulent drug effect. Without such a correction for reagent variation, assurance is lacking, that the prothrombin time ratio is reproducible with different reagents, from different test methods, or in different laboratories. Indeed, before the introduction of the ISI reagent sensitivity measure and the INR reporting convention, anticoagulant regimen in North American patients with AF tended to produce more anticoagulation than was necessary, increasing bleeding risk without adding antithrombotic benefit, because less sensitive thromboplastins had been put into increased use over the previous two decades. After the introduction of the INR, four of the randomized controlled comparable studies, mentioned above, determined a level of oral anticoagulation that was maximally safe and effective, to prevent stroke while avoiding drug induced bleeding. This target zone for anticoagulation of patients with AF is represented by an INR between 2 and 3.

Laboratory Practice Collaboration: the challenge

To grasp this opportunity, laboratorians need to help clinicians overcome two difficulties that were identified by the National Committee for Quality Assurance (NCQA), the accrediting body for health insurance plans, when the NCQA considered warfarin treatment in AF patients as a quality measure in the “domain” of “effectiveness of care” for treating chronic illness, in the NCQA’s Healthcare Employers’ Data Information Set (HEDIS). The NCQA wanted to include warfarin treatment of AF patients in HEDIS on the following basis: (1) the level of benefit from warfarin prophylaxis, demonstrated in controlled studies, was not being reached in clinical practice; (2) the shortfall was due to underuse of a proven intervention; (3) this underuse missed an opportunity to prevent stroke, a major foreseeable complication of AF; (4) as a complication, stroke has relatively high mortality and morbidity and, beyond these human costs (5) substantial health care dollars could be saved, if AF patients without contraindications to warfarin therapy were treated by oral anticoagulation to maintain INRs between 2 and 3. In considering the monitor, the NCQA’s Committee on Performance Measurement found two key problems: (1) absence of convenient ways to ascertain whether AF patients had contraindications to warfarin prophylaxis and (2) need for evidence that the intervention was achieving its intended effect. These problems, in the Committee’s 1996 view, precluded adoption of the AF-anticoagulation monitor by HEDIS.

So the situation has remained: the appropriate INR target range has been located, reaching it in clinical practice remains a challenge. In one study from the mid-1990’s, the appropriate 2-3 INR range was reached during only 50% of monitored patient days, in 30% of days the level of anticoagulation was above the 2-3 range and below the range in the remaining 20%.

Laboratory Practice Collaboration: the obstacles

From a laboratory standpoint solving these two problems requires clinical information. Some of the necessary information is relatively available to laboratories, but sources for other equally critical information are not easily discovered.

Relatively available data includes the diagnosis of AF itself, as the indication for testing: the ICD-9 code for AF (427.31) is both reproducible and frequently cited in orders for PTs. Other diagnoses that increase the risk of stroke, hence the importance of oral anticoagulation for AF, may also be available as laboratory test indications: hypertension, congestive heart failure, coronary artery disease, diabetes mellitus, previous stroke, and myocardial infarction. This is however less likely to be true for salient contraindications to oral anticoagulation: previous gastrointestinal or intracranial hemorrhage and renal insufficiency. The other major contraindication, a history of falls prompting medical attention, as well as the most common relative contraindication, the diagnosis of dementia, are even harder to come by in the laboratory setting. Information easily elicited by the laboratory, however, does include valuable data for stratifyng AF stroke risk: patient age, the most important modifying variable for the risk of AF, and sometimes whether or not AF patients are followed in specialized oral coagulation clinics. These clinics appear, in some published studies, to increase the efficiency of warfarin treatment in practice, decreasing stroke risk in AF.

The proportions of INR’s ‘below’, ‘in’ or ‘above’ the 2-3 INR ranges that is, inadequate, appropriate, or excessive drug effect, as well as pertinent stratifiers, which put that key index in a clinical context, are not part of the routine laboratory record. These stratifiers include documentation of duration of therapy, recent alteration of the regimen, and the frequency of PT monitoring. Duration is an important variable, because of the expectation that, in patients on oral anticoagulation for more than 6 months, INRs should be ‘stable’ [that is, in the 2-3 INR range more than two-thirds of the times]. This expectation is often not satisfied among patients newly started on a warfarin regimen, so failure to demonstrate stability is regarded as more understandable in this new patient subgroup [patients during the first 6 months of treatment]. Documentation of recent alteration in the prophylactic regimen registers another modifying variable that decreases the expectation of stability. A final important modifying variable, that may or may not be easily available to the clinical laboratory, is whether or not the frequency of prothrombin time monitoring of oral anticoagulation has changed recently: increased frequency of testing is associated with unstable anticoagulation regimes, while decreased and consistent frequency of monitoring events is associated with stable regimens.

Laboratory Practice Collaboration: the objective

Collaboration between clinical laboratory and clinician’s office needs to have its focus clearly understood by both groups: the objective is increasing the fraction of AF patients receiving warfarin (i) who have no contraindications to oral anticoagulation (more than 80% of all patients with AF), (ii) who have INR’s in the 2-3 therapeutic range, (so are being treated at the best benefit: risk ratio), and (iii) who have demonstrated the stability of such therapeutic effect, where stability of an oral anticoagulation regimen is defined as at least two-thirds of INR’s in the 2-3 INR range over a three month period.

Laboratory Practice Collaboration: the solution

The information just summarized, that is not immediately available to the laboratorian, is usually only a telephone call away. It resides in the offices of the physicians caring and prescribing for the patients with AF. Most warfarin regimens are managed by telephone calls: (1) calls from laboratories with INR results to the prescribing physician’s practice and (2) calls from the practice office to patients with directions for continuing or changing the oral anticoagulation regimen. The first group of incoming calls are usually received and the second group of outgoing calls are usually made by the prescribing physician’s office or clinical staff. These staff members are guided by formal or informal algorithms developed and overseen by the prescribing physician to connect the laboratory results with their therapeutic implications. When asked to collaborate with the clinical laboratory in an effort to improve oral anticoagulation’s effect, cooperating office staff can usually provide documentation of the presence or absence of contraindications; they almost always can report the duration of a regimen, and know whether it has been recently changed.

Making the Collaboration Work

The laboratory can begin the collaboration by identifying practices with evaluable patients: (a) reviewing prolonged prothrombin times, (b) culling out prolonged PTs belonging to patients identified by indication as suffering from AF, and (c) flagging these patients so that their PTs over the next 3 months will be collated by the participating laboratory. The collaborating practices then provide the following information when the study period is complete (1) duration of therapy, (2) frequency with which INRs were obtained over the study period,(3) frequency with which the regimen was changed during that period, (4) frequency with which the warfarin regimen was managed using, (5) anticoagulant clinic or (6) home assessment program.

Outcome Measures

Analyzing the data assembled in this way yields four indices of clinical outcome. The four indices are: (a) “effectiveness” (the HEDIS proposed measure)-AF patients with the most recent INR (the last INR prospectively observed) in the 2-3 range/all INRs observed in warfarin anticoagulated AF patients; (b) “stability”-AF patients with two-thirds of INRs over the study periord in the 2-3 range; (c) “under anticoagulation”- AF patients with the most recent INRs < 1.5/AF patients with most recent INRs. (d) “over anticoagulation”-AF patients with most recent INRs > 3.5/AF patients with most recent INRs. These rates can be stratified by the following six variables characterizing patient populations: (1) patient age, grouped by decades, (2) fraction of patients on oral anticoagulation less vs. more than 6 months, (3) frequency of prothrombin time testing of less than vs. more than once/monthly during the study period, (4) frequency of warfarin regimen change of less than or equal to once during the study period vs. regimen change more than once, (5) presence vs. absence of anticoagulation clinic monitoring, (6) and presence vs. absence of home monitoring.

Study Limitations

Like all efforts to link laboratory information to measurable outcomes, this collaboration has some limitations. Although about 30% of PTs performed in hospital laboratories are for monitoring warfarin therapy, many prothrombin times from AF patients are now generated either by non-hospital laboratories with little access to patient data or from hand-held devices brought to the patient’s home. For this reason, the participating laboratory’s set of INRs may not represent all the results that went into management decisions about the patient population during the study period. Similarly, although AF is a condition of growing prevalence, an individual practice would have to follow about 10 AF patients over a 3-4 month period to generate reproducible practice-specific

“effectiveness”, “stability”, “under”-, and “over anticoagulation” rates. When the denominators of the outcomes measures are assembled from multiple practices, there is always the possibility that hidden differences among practices will reduce the measure’s validity for measuring the outcomes of a specific practice.

Conclusions

Adequate, safe warfarin anticoagulation in AF patients is an important, growing clinical task. Laboratory monitoring of prothrombin time ratios, expressed as INRs, is the key to achieving this intervention’s effect on outcome. This critical outcome is prevention of stroke, without increased risk of major or intracerebral bleeding.

Collaboration between a clinical laboratory and warfarin-prescribing office practices can produce a demonstrably effective monitoring program. This is an important instance of the laboratory collaborating with clinicians to measure the impact of a laboratory test on the outcomes of a critical therapeutic intervention.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download