STOPPING LONG TERM ASPIRIN INCREASE CARDIOVASCULAR EVENTS

Patients who discontinue long-term, low-dose aspirin therapy for reasons other than surgery or bleeding face an increased risk of cardiovascular events, a large population-based study suggests^[1].

After a median 3 years of follow-up, patients who stopped taking aspirin had a 37% higher rate of hospitalization for MI, stroke, or cardiovascular death (adjusted hazard ratio 1.37; 95% CI 1.34–1.41).

This translated into an absolute risk increase of 13.5 events per 1000 person-years at risk. Simply put, one of every 74 patients who discontinued aspirin had an additional CV event in 1 year.

Stopping aspirin was especially perilous for patients with previous CVD, with one of every 36 secondary-prevention patients experiencing an additional CV event per year compared with one of every 146 primary-prevention patients who quit aspirin, according to study authors Drs Johan Sundström and Jonas Oldgren (Uppsala University, Sweden) and colleagues.

"Millions of patients worldwide take aspirin on a daily basis and might consider stopping at some time during their life. We performed this study to help physicians and patients to make an informed decision whether or not to stop aspirin use," Sundström told theheart.org | Medscape Cardiology.

Discontinuation rates of about 10% to 20% have been reported in patients with a recent MI and reached up to 30% in broader patient settings, despite strong evidence and guideline recommendations for low-dose aspirin in secondary prevention and potential utility in primary prevention. The public-health effects of discontinuing long-term aspirin, however, are not well known, the authors note in their report, published September 26, 2017 in Circulation.

To examine this issue, the investigators took advantage of the fact that low-dose aspirin is sold only by prescription in Sweden and linked the Swedish prescription drug register with mandatory inpatient and cause-of-death registers for 601,527 patients on low-dose aspirin for primary or secondary prevention between July 2005 and December 2009. Because major bleeding and surgical procedures can affect thrombogenicity and cause aspirin discontinuation, the first 3 months after either event were excluded from the time at risk.

Half of the sample was female, the mean age was 73 years, 54% used aspirin for secondary prevention, and half were on long-term aspirin without a prior hospitalization for CVD.

About 15% of long-term aspirin users were off treatment after 3 years and about 20% of aspirin-naïve patients did not collect a second aspirin prescription, although the reasons for stopping treatment are not known, Sundström said.

During follow-up, 62,690 CV events occurred and 73,636 patients died.

Being off aspirin was associated with an increased risk of CV events in patients using aspirin for primary prevention (HR 1.28; 95% CI 1.22–1.34) and secondary prevention (HR 1.46; 95% CI 1.41–1.51).

Patients who were older and had prior CVD were at higher elevated risk for CV events when off aspirin, while treatment with other antiplatelet or oral anticoagulant drugs was associated with a lower CV event-risk when off aspirin.

Importantly, "the [elevated CV] risk increased shortly after discontinuation and did not appear to diminish over time; hence, adherence to low-dose aspirin treatment in the absence of major surgery or bleeding is likely an important treatment goal," Sundström said.

He pointed out that experimental studies have suggested a "rebound effect" a few days after stopping aspirin, when concentrations of aspirin in the blood are very low, and blood clotting is stimulated rather than inhibited. "The clinical meaning of this was hitherto unknown."

While it's unknown whether or when the rebound effect occurs, the investigators suggest the clinical importance of a rebound effect "may be substantial given the large number of aspirin users and the high discontinuation rates."

The possibility of such mechanisms is also supported by the finding that aspirin discontinuation was not associated with CV events in patients protected by other antiplatelet or oral anticoagulant drugs, although they were likely at higher absolute risk for such events.

Finally, limitations of the study include the possibility of confounding and a lack of data on socioeconomic status or physical examinations including blood pressure and lipids or lifestyle measures like smoking.

A similar but weaker association with the secondary outcome of nonfatal CV events—increased 10% among people off vs on aspirin—may also signify some reverse causation, the authors write.

The study was partly funded by AstraZeneca by funds to Statisticon for statistical analyses. Sundström reports serving on an advisory board for AstraZeneca. Disclosures for the coauthors are listed in the paper.