Dementia Risk in AF Rises With Worsening Warfarin Management: Cohort Study

August 14, 2014

MURRAY, UT — The better warfarin anticoagulation is managed over the long term in patients with atrial fibrillation (AF), the lower their later risk of dementia, suggests a retrospective single-center study[1]. It found an inverse relationship between the percentage of time in the therapeutic range (TTR) of INR and long-term risk of a dementia diagnosis in AF patients initially without either dementia or a history of stroke.

"Our results suggest that it is not necessarily anticoagulation exposure per se, but the degree of exposure to over- and/or underanticoagulation that confers risk of dementia," according to researchers.

That, they say, supports the possibility that chronic overanticoagulation and underanticoagulation can produce repeated microemboli and microbleeds that may be below detection thresholds of standard brain imaging but eventually contribute to dementia.

The study, based on 2605 patients tracked an average of four years up but up to nearly 10 years, was published August 8, 2014 in Heart Rhythm with lead author Victoria Jacobs (Intermountain Medical Center, Murray, UT).

Most cases of dementia are diagnosed as Alzheimer's disease, and this was true in the current analysis as well, notes the study's senior author, Dr T Jared Bunch (Intermountain Medical Center). The findings, he told heartwire , support a line of research pointing to a vascular component in the pathophysiology of Alzheimer's.

It may be better to use a more predictable anticoagulation strategy in AF patients with chronically poor INR control, he said, or perhaps switch them to one of the new oral anticoagulants (NOACs) on the possibility that they will provide a more consistent effect.

It isn't known whether NOACs improve cognition, of course, and while the evidence says they can cut the risk of larger bleeds, "we don't know if they reduce microbleeds or microemboli. So that would have to be teased out prospectively," he said.

"But I think until we know that, [the NOACs] would probably be a safer choice compared with trying to maintain a patient [on warfarin] who consistently has a low percentage of time in therapeutic range."

The group looked at 2605 patients with no history of dementia or cerebrovascular events who were chronically managed on warfarin with a target INR of 2 to 3. Their mean age was 74, and 54% were male; 30.5% had a CHADS2 score of 0 or 1, and about 70% had a CHADS2 score of 2 or 3.

Over a median of four years (maximum 9.9 years), the percent TTR averaged 63.1%; mean time in INR <2 was 25.6% and mean time in INR >3 was 16.2%. The rate of a dementia diagnosis was 4.2%, mostly Alzheimer's dementia (2.5% overall) but also "senile" (1.4% overall) and "vascular" (0.3% overall) dementia.

The risk of dementia went up 1.7% for each percentage-point increase in time with an INR <2, the group writes, and by 1.8% for each percentage-point increase in time with an INR >3 (p=0.005).

The incidence of a dementia diagnosis fell by increasing TTR quartile category, as did the multivariate adjusted hazard ratio for a dementia diagnosis in the first through third TTR quartile categories vs the highest one, which was >75% of time in TTR.

Incidence (%) of Dementia and Adjusted Hazard Ratio (HR) for Dementia by TTR Quartile Groups

Time in INR therapeutic range (%) Incidence HR, p p
<25 5.8 5.34 <0.0001
26–50 5.5 4.10 <0.0001
51–75 4.9 2.57 0.001
>75 1.9* index for HR
*p=0.001 for incidence trend

Although an observational study can't discern cause and effect, and therefore the findings could reflect that patients with dementia often do not maintain good INR control, "we think there's something more to it than that," Bunch said. About 25% of patients on warfarin maintain poor control not necessarily because of fewer INR checks but likely based on the genetics behind how they metabolize the anticoagulant, he speculated, "and that group may be at particular risk for dementia."

All prospective long-term NOAC trials should look at any possible associations with cognitive function "and include those in their analyses, so we can understand the impact of the newer therapies on dementia risk."

Jacobs had no disclosures. Bunch discloses serving on an advisory board for Boston Scientific. Disclosures for the coauthors are listed in the article.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.