Stretching the INR Recall Interval
For Some, Less Testing Might be Better
By Bill Malone
Current guidelines, based on expert opinion, recommend an interval of no more than 4 weeks between INR monitoring for patients taking anticoagulants such as warfarin. More frequent testing has been suggested for less stable patients; however, for those with long-term INR stability, researchers are considering whether intervals longer than 4 weeks might be sufficient, or even beneficial. A new study described in this issue of Strategies looked at a large cohort of anticoagulated patients to establish predictors of INR stability.
Many studies have focused on the best international normalized ratio (INR) target range for anticoagulation, but few examine testing frequency, leaving clinicians with no evidence-based guidelines for how often patients should be tested. Because anticoagulants are infamously unpredictable and operate within a narrow therapeutic index, it has seemed logical to err on the side of more testing rather than too little.
However, recent research suggests that for some very stable patients, longer intervals between testing may be associated with improved INR control, due to a reduction in anticoagulant drug related adverse events. Now, for the first time, a new study has looked at a large retrospective cohort of patients taking anticoagulants to quantify and describe a group of patients for whom frequent testing might not be necessary or helpful (Blood 2009;114:952–956).
The aim of the research was to lay the groundwork for prospective studies that can accurately pinpoint those patients who might benefit from less frequent testing. “Based on clinical observations, it was apparent to us that there was a subfraction of our anticoagulated population—we didn’t know exactly how large—that had very good INR control, meaning that you almost never had to change their warfarin dose,” said Daniel Witt, PharmD, FCCP, senior manager of the clinical pharmacy anticoagulation service of Kaiser Permanente Colorado in Lafayette, Colo., and an author of the study. “Most significantly in this study, we were able to identify a fairly large number of individuals within our population that had really good, stable INR control, giving some glimpse of who are going to be those very stable patients.”
Witt and his colleagues drew insight from the United Kingdom, where anticoagulation providers routinely monitor INR in longer 8 or 12 week intervals. In fact, the software (Dawn AC-4S Systems, Ltd, Cumbria, UK) Witt uses to manage Kaiser’s large anticoagulation service comes from a U.K.-based company that also offers clients data analysis. The company takes anonymized data that Witt sends in for benchmarking twice a year and compares the therapeutic control of Witt’s anticoagulation service to other sites around the world that use the same software. The company performed regression analysis for the Kaiser group and turned up some interesting results. “They were able to identify that if someone was out of range and we changed the warfarin dose, it was best if we brought them back within a week or so. But when people were in range, if we lengthened out the time between the INR interval, it seemed to improve the time in range,” said Witt.
One reason proposed for this outcome—a finding researchers have noticed elsewhere—is that the more often patients are tested, the greater the signal to noise variability, meaning a clinician or pharmacy might overreact to a relatively subtle INR change, explained Witt. “The more frequently you test, the more likely you are to catch a stable patient who’s just outside of their INR range, and then you may feel compelled to respond and change a dose.” Longer intervals between testing could avoid these overcorrections.
This scenario is the inverse of how patients often use fingerstick INR testing. Though they test every week, many times they don’t take action on every odd result, knowing from experience that the next week they test, they’ll likely be back in range. “Perhaps instead of testing frequently to reassure yourself that everything is okay, what we’re seeing is that if you have someone who is stable most of the time, it’s probably best not to fool around with it. Check the INR less frequently and then you feel less compelled to change the dose,” said Witt.
For the study’s analysis, Witt and his colleagues drew from Kaiser Permanente Colorado’s electronic medical records that integrate pharmacy and laboratory data. The team started with 7,686 patients, and whittled down the number to 6,073 patients who had a period where INR was measured regularly for at least 6 months. Stable patients were those who achieved 100% INR control, with all of their INR values within a strict therapeutic range during the period. This definition netted 2,504 stable patients and 3,569 comparator patients.
Digging into the data further, a profile emerged of these stable patients: more than 70 years of age and the absence of comorbid diabetes, heart failure, or concurrent estrogen therapy. They also tended to have an INR target of less than 3.0 and have low chronic disease scores. While the rate of overall mortality was lower in the stable group, the anticoagulation-therapy-related mortality rate was not significantly different between stable and comparator patients. The authors note that their finding that advanced age predicted long-term INR stability is “counterintuitive and should be confirmed in additional studies.” There are several plausible explanations for the age factor. Witt speculated that it might be due to the simple stability of an older person’s daily routine. Also, younger people are often on anticoagulants for different reasons than elderly patients.
Witt also pointed out the part of the paper’s strength could also limit the transferability of the results. Due to Kaiser’s large clinical pharmacy anticoagulation service, the data are comprehensive, with all INR testing performed by a single lab and collected systematically in an electronic medical record. However, for patients who aren’t part of such a coordinated, all-inclusive system, clinicians may not feel comfortable letting them go for long periods in between testing, especially when there isn’t a good system of picking up people who don’t return for INR monitoring when they’re supposed to. Meanwhile, Witt expects researchers to build on this paper with randomized trials that could built a more complete profile of who needs intensive INR monitoring and who doesn’t.