68-year-old female with a history of hypertension, diabetes mellitus, stroke, and atrial fibrillation presents for routine follow-up at your hospital’s anticoagulation clinic. The clinical pharmacist checks her international normalized ratio (INR) with a point-of-care (POC) device to monitor her anticoagulant therapy (warfarin). The POC INR result is elevated to 4.0, which is above the recommended INR goal range of 2.0–3.0 based on her clinical indication of atrial fibrillation. The pharmacist enters the POC INR result into the patient’s electronic medical record and discovers she had an INR obtained earlier that same day, along with a basic metabolic panel and complete blood count ordered by her primary care physician. The clinical lab (CL) INR result was 2.9, and obtained just 90 minutes earlier. All other test results were normal.
The patient says she hasn’t started any new medications or had recent warfarin dose changes, nor has she experienced bleeding symptoms or changed her diet. INR measurements from the past 6 months have all been within range. Given the conflicting INR information, the pharmacist is uncertain whether the patient should continue her current warfarin dosing schedule or have an adjustment. Why are the INR results so different, and which result should be used to guide warfarin dosing recommendations?
Scenarios like this challenge physicians, pharmacists, and laboratory medicine professionals tasked with managing anticoagulation therapy. This review highlights some of the benefits and limitations associated with POC INR testing as well as the discordance between POC and CL INR assays, and suggests strategies for managing discordant INR values.
Background
Physicians worldwide prescribe vitamin K antagonists (VKAs) for chronic anticoagulant therapy across a number of clinical indications. In the United States, warfarin is the predominant VKA used for venous thromboembolism treatment and prophylaxis against stroke and systemic embolism in patients with atrial fibrillation or mechanical heart valves.
However, numerous patient-specific factors influence warfarin sensitivity. These include age, body mass, nutritional status, hepatic function, and genetic variations in the cytochrome P-450 complex (CYP2C9) and vitamin K epoxide reductase complex 1 (VKORC1) (1). In addition to these patient-specific factors, warfarin has a narrow therapeutic window that requires frequent INR monitoring to minimize hemorrhagic and thromboembolic complications.
POC INR Testing
The INR method was developed to standardize prothrombin time (PT) results between different PT analyzers and thromboplastin reagents (1). Over the past 2 decades manufacturers have engineered several POC INR devices that use fingerstick samples of whole blood instead of platelet poor plasma to measure clotting time. These instruments then estimate INR using conversion formulas derived by the device manufacturers from comparison of clotting times to a reference PT assay. Potential advantages of POC INR testing include rapid turnaround times (typically within seconds) and ease of frequent out-of-hospital testing. These attributes greatly facilitate INR monitoring in home-bound or transportation-challenged patients, as well as in those for whom drawing venous blood samples is difficult. Furthermore, POC INR devices enable some patients to self-test and self-manage their warfarin therapy with favorable outcomes (2). Anticoagulation management services also have reported that POC INR devices improve clinical efficiency and patient comfort and satisfaction compared to traditional venipuncture methods (3).
Despite these benefits, POC INR testing has several disadvantages. These are due to result interference from severe anemia or polycythemia (hematocrit below approximately 25% or above about 55%), co-administration of other anticoagulants with warfarin (e.g., low molecular-weight heparin), fibrinogen level, and anti-phospholipid inhibitors. Even in the absence of these known interfering conditions, significant discrepancies between CL INR and POC INR values are common, especially when CL INR >3.0 (4,5). Commercially available POC INR devices typically use either an electrochemical sensor or a mechanical method to determine the test endpoint, thereby estimating the INR. A comparison between two POC INR devices using similar test methodology and a reference lab INR found significant differences between POC INR results, suggesting that the observed discrepancies between POC INR devices are not solely attributable to differences in test methodology (e.g., mechanical vs. electrochemical) (6). In addition, the type of thromboplastin used in the CL INR assay can influence the degree of INR discordance between methods.
These results often are sufficiently different to alter clinical decisionmaking related to anticoagulant dose adjustments (4-6). The Food and Drug Administration’s (FDA) recent recall of a POC INR device (INRatio) used in the pivotal ROCKET-AF trial highlighted this problem. FDA issued the recall due to concerns of potential harm associated with warfarin dosing based on POC INR results that were “clinically significantly lower than those found by a laboratory method” (7). Stemming from this controversy, FDA held a public workshop exploring the problems affecting POC INR and CL INR correlation, without clear solutions. Notably, an international standard published in 2007 suggested that 90% of POC INR values should correlate within 30% of CL INR values between 2.0-4.5. Although FDA has not adopted this standard, more stringent correlation standards were recently proposed (≥95% of POC INRs should correlate within ±20% of CL INR) (11).
POC INR Confirmation Protocols
The complex interplay between patient-specific factors, dietary intake of vitamin K, drug-drug interactions, and the effect of acute illness on warfarin metabolism frequently results in supratherapeutic INR values. The magnitude of INR elevation not only correlates with the risk of bleeding, but also directly influences clinical decisionmaking. For example, if an INR result is above the therapeutic range but <5.0, guidelines recommend holding the next dose of warfarin and/or adjusting the maintenance dose. However, when the INR rests between 5.0-9.0 and the patient has no signs of bleeding, warfarin therapy might be curtailed for 1-2 doses and a low-dose of vitamin K considered. When the INR is >9 and the patient does not have signs of bleeding, warfarin typically would be stopped and vitamin K administered. Administering vitamin K under these circumstances may result in prolonged periods of subtherapeutic INRs, increasing patients’ thrombotic risk. Similarly, frequent warfarin dose adjustments may result in unstable INRs, a known risk factor for mortality (8). Given these potential adverse consequences, avoiding both unnecessary warfarin dose adjustments and vitamin K administration remain guiding principles of anticoagulation management. Patients should be encouraged to monitor their INR at a consistent cavesite, thereby avoiding unstable INRs introduced when different locations use different monitoring devices or reagents.
High-quality anticoagulation management services typically know these risks and have standard operating procedures to confirm elevated POC INR results and monitor agreement between CL INR and POC INR. For example, at our institution we confirm with CL INR all POC INR results >5.0, and base all warfarin adjustments on the CL INR value. In addition, we perform annual CL INR to POC INR correlation studies, as well as competency testing for healthcare providers performing POC INR testing. These quality assurance measures heightened our awareness of INR discordance within our institution and led us to develop a POC INR correction factor.
POC INR Correction Factors
In spite of efforts to harmonize PT assays with the advent of the INR, inter-laboratory variability likely will persist until the medical community reaches consensus on a standardized thromboplastin preparation. Although most institutions consider the CL INR as the “gold standard” or reference standard, the absence of large clinical trials demonstrating superiority of one INR method over the other adds to the controversy. Until further standards or clinical trials become available, CL INR and POC INR discrepancies will remain, leaving healthcare providers wondering which to believe.
In an attempt to remove some of this confusion and harmonize INR results, some anticoagulation management services have derived correction factors that can be applied to POC INR values to predict the CL INR value (9,10). These correction factors seek to minimize the difference between POC INR and CL INR results and clarify subsequent clinical decisions. For example, within our institution applying a correction factor to POC INRs >3.0 improved agreement to the CL INR (within ±15%) in more than 70% of tests (9). A Bland-Altman plot comparing uncorrected (Figure 1) and corrected POC INRs highlights the effectiveness of this approach (Figure 2). Applying this correlation factor in our institution reduced differences in clinical decisionmaking by 43%.
These findings have altered our clinical approach to POC INRs, and for values between 3.0-5.0, we routinely apply a correction factor to estimate the CL INR. Similarly, Richter et al. demonstrated the feasibility of applying a POC INR correction factor to guide warfarin dosing among patients with supratherapeutic POC INRs (>4.0), compared to confirmatory venipuncture samples. The differences in warfarin dosing decisions were negligible and the corrected POC INR approach was more efficient (10).
The methodology used to derive a correction factor is simple (linear regression) and has been published elsewhere (9,10). POC INR devices may overestimate or underestimate the CL INR. This means that correction factors are device- and institution-specific and cannot be directly applied across institutions with different POC devices and lab analyzers. In addition, changes in reagent lots, lab analyzers, POC devices, or test strip lots may influence the correction factor equation. Laboratories should recalculate their correlation factor every 6-12 months and following any significant changes in lab equipment.
Clinical Case Continued
The patient described at the beginning of this review has multiple stroke risk factors in the setting of atrial fibrillation. Based on the stroke risk prediction scoring system, called CHA2DS2-VaSc, she has 6 points, placing her in the moderate-high risk group with an approximate 9.7% annual risk of stroke. The pharmacist is appropriately concerned about the patient’s elevated POC INR of 4.0, and must balance the small increased risk of bleeding with the patient’s underlying stroke risk. An inappropriate warfarin dose reduction could result in a subtherapeutic INR and inadequate protection from a stroke or systemic embolism. The additional history obtained from the patient did not uncover any new medications that would interfere with the POC INR device (e.g., low molecular-weight heparin). The patient’s hematocrit was normal at 39% and no phospholipid antibodies were detected during her prior stroke evaluation.
While the patient’s POC INR is elevated, the CL INR is within the therapeutic range. A local policy is in place requiring POC INRs >5.0 to be confirmed with a venipuncture sample and CL INR. However, the patient’s INR is not above this threshold and the pharmacist remains perplexed as to which value to believe. Ultimately, the pharmacist leaves the warfarin dose unchanged and instructs the patient to return in 1 week for a follow-up INR. At this follow-up visit, the patient’s POC INR is 2.9. Although she did not suffer any harm, had a POC INR correction factor been available and applied, the patient’s inconvenience and cost of follow-up INR could have been avoided.
Conclusion
POC INR testing for chronic anticoagulation monitoring enables patient self-testing and self-monitoring, and is more convenient and efficient than CL INR testing. POC INR testing is widely used in outpatient labs and anticoagulation clinics, although CL INR testing remains the reference standard. Laboratory professionals should be aware of the potential for POC INR and CL INR disagreement and work to develop institutional procedures for confirmatory testing. POC INR correction factors are promising tools to improve INR agreement and clinical decisionmaking.
References
1. Pendleton RC, Rodgers GM. Disorders of hemostasis and coagulation. In: Greer JP, Arber DA, Glader B, List AF, Means RT, Paraskevas F, Rodgers GM, editors. Wintrobe’s Clinical Hematology. 13th Ed. Lippincott Williams & Wilkens; 2014. p. 1241-1243.
2. Heneghan C, Ward A, Perera R, et al. Self-monitoring of oral anticoagulation: Systematic review and meta-analysis of individual patient data. Lancet 2012;379:322-4.
3. Chaudhry R, Scheitel SM, Stroebel RJ, et al. Patient satisfaction with point-of-care international normalized ratio testing and counseling in a community internal medicine practice. Manag Care Interface 2004;17:44-6.
4. Hur M, Kim H, Park CM, et al. Comparison of international normalized ratio measurement between CoaguChek XS Plus and STA-R coagulation analyzers. Biomed Res Int 2013;2013:213109.
5. Moore GW, Henley A, Cotton SS, et al. Clinically significant differences between point-of-care analysers and a standard analyser for monitoring the International Normalized Ratio in oral anticoagulant therapy: a multi-instrument evaluation in a hospital outpatient setting. Blood Coagul Fibrinolysis 2007;18:287-92.
6. Donaldson M, Sullivan J, Norbeck A. Comparison of International Normalized Ratios provided by two point-of-care devices and laboratory-based venipuncture in a pharmacist-managed anticoagulation clinic. Am J Health Syst Pharm 2010;67:1616-22.
7. Cohen D. Manufacturer failed to disclose faulty device in rivaroxaban trial. BMJ 2016;354:i5131.
8. Van Den Ham HA, Klungel OH, Leufkens HG, et al. The patterns of anticoagulation control and the risk of stroke, bleeding and mortality in patients with non-valvular atrial fibrillation. J Thromb Haemost 2013;11:107-15.
9. Johnson SA, Vazquez SR, Fleming R, et al. Correction factor to improve agreement between point-of-care and laboratory International Normalized Ratio values. Am J Health Syst Pharm 2017;74:e24-31.
10. Richter C, Taylor J, Shuster J. Correction of point-of-care INR results in warfarin patients. Point Care 2016;15:1-3.
11. Goehe RR, Riddick K. FDA regulatory oversight of POC PT/INR in vitro diagnostic devices. 2016; Public workshop - Point of care prothrombin time/international normalized ratio devices for monitoring warfarin therapy. http://www.fda.gov/MedicalDevices/NewsEvents/WorkshopsConferences/ucm476561.htm. (Accessed October 15, 2016).
Stacy A. Johnson, MD, is the medical director of the University of Utah Health Care thrombosis service and an assistant professor of medicine in the division of general internal medicine at the University of Utah School of Medicine. +Email: [email protected]