Conference Coverage

Warfarin boosts OA risk in Rotterdam Study



TORONTO – The use of warfarin or related vitamin K antagonists was associated with a more than 100% increased risk of incident or progressive knee or hip osteoarthritis in the Rotterdam Study, Cindy G. Boer reported at the OARSI 2019 World Congress.

A biologically plausible mechanism exists for this relationship, added Ms. Boer, a PhD student with a special interest in the molecular genetics of OA at Erasmus University, Rotterdam, the Netherlands.

Cindy G. Boer of Erasmus University, Netherlands Bruce Jancin/MDedge News

Cindy G. Boer

In a previous genetic study, she and her coinvestigators identified two genetic variants that result in loss of function of matrix Gla protein (MGP), a key inhibitor of cartilage calcification. They showed that the presence of these alleles was associated with a significantly increased risk of hand OA, which makes sense because increased calcification within a vulnerable joint promotes OA.

“The interesting thing here is that, in order for MGP to inhibit calcification, it needs to be gamma-carboxylated by vitamin K. If it’s not gamma-carboxylated it cannot inhibit calcification,” Ms. Boer said at the meeting sponsored by the Osteoarthritis Research Society International.

This observation led her to hypothesize that patients on warfarin or other vitamin K antagonists might have an increased risk of developing new-onset OA or, if they already had OA, of experiencing disease progression, since their medication inhibits MGP. To test this hypothesis, she and her coinvestigators turned to the landmark Rotterdam Study, a prospective, population-based cohort study of 15,000 participants, ongoing since 1990. Two large cohorts within the study had data available on the incidence and progression of radiographic knee and hip OA, one group over the course of 5 years of follow-up, the other with 10 years.

Serial x-rays of 8,845 knee joints were available, including 657 of warfarin users. Their rate of incident or progressive knee OA was 13%, compared with 5.9% in the nonusers in an analysis adjusted for age, sex, BMI, baseline OA status, and time between study visits.

In a similar vein, the rate of incident or progressive hip OA was 12% in the warfarin users, compared with 3.1% in nonusers.

About 80% of the OA endpoints in this analysis involved incident OA, defined radiographically as Kellgren-Lawrence grade 2. Progressive OA was defined as going from grade 2 at baseline to grade 3-4 during follow-up.

There was a signal of a treatment duration-related effect, with OA rates trending highest in individuals on warfarin for longer than 365 days, followed by those on the oral anticoagulant for more than 100 days, who in turn had higher rates than those on warfarin for less time.

Ms. Boer said an important next step in this research is to replicate the warfarin/OA association in an independent cohort. Also, she and her coworkers are now gathering OA incidence and progression data in patients on direct oral anticoagulants rather than warfarin to test the hypothesis that they will not have an increased rate of OA, compared with nonusers, since these newer agents don’t affect vitamin K. Of course, if they do turn out to have an elevated risk, it would point to one or more of the conditions for which oral anticoagulants are commonly prescribed as the explanation.

She reported having no financial conflicts regarding her study, sponsored by Erasmus University and the Dutch government.

Next Article: