June 14, 2017
Taking aspirin for secondary prevention of stroke or myocardial infarction (MI) is associated with a higher than expected risk for disabling or fatal bleeding in people aged 75 years and older, a new observational study shows.
The authors suggest that all patients aged 75 years or older prescribed aspirin for secondary prevention should also be given a proton-pump inhibitor (PPI) to protect against gastrointestinal (GI) bleeding.
The study was published online in The Lancet on June 13.
"We knew before that there was an increased risk of bleeding with aspirin in the elderly, but what we didn't know was how high that risk was and the consequences of those bleeds," senior author, Peter Rothwell, MD, John Radcliffe Hospital, University of Oxford, United Kingdom, explained to Medscape Medical News.
"There has been a sense that antiplatelets prevent important ischemic events, such as stroke and MI, and while we knew there is an increased risk of bleeding, this was seen as bit of a nuisance and not equivalent to the ischemic events. But our data shows that actually the bleeding is more likely to be disabling than the strokes in this age group so really needs to be taken more seriously."
Professor Rothwell notes that many of these bleeds are preventable with PPIs.
"About half the bleeds in this study were GI bleeds and more than half of disabling or fatal bleeding were GI, and we know from previous studies that PPIs can prevent about 80% of GI bleeding caused by antiplatelet agents."
He pointed out that current guidelines advise that patients receiving long-term antiplatelet drugs should be co-prescribed a PPI if they are high risk for bleeding, but they don't define "high risk."
"Our data show that pretty much anyone 75 years of age or over should be categorized as high risk," he said. "I think we could make a reasonable argument for everyone in this age group who is taking long-term antiplatelet therapy for secondary prevention of MI or stroke to be also prescribed a PPI."
Is Aspirin Necessary Long Term?
But the question has also been raised as to whether long-term aspirin should be prescribed for these patients in the first place.
Commenting for Medscape Medical News on the new study, John Cleland, MD, Imperial College London, United Kingdom, who has long been an aspirin skeptic, suggested that these new data provide further evidence against giving aspirin long term after a stroke or MI.
"There is no compelling evidence that aspirin continued for greater than 28 days after an event is beneficial," Dr Cleland said. "The antiplatelet meta-analysis is based on long-term trials that were neutral and done more than 40 years ago before the era of modern medicine. ISIS-2, the only truly positive study, only treated post-MI for 28 days."
"Doctors should know they are using guesswork rather than evidence-based medicine if they give aspirin for longer than 28 days," he added. "We need to stop prescribing useless polypharmacy. And PPIs can have risks too, such as increased risk of Clostridium."
Responding to this, Professor Rothwell said, "It's a very difficult question on whether aspirin should be given to older patients long term as all the studies have been done in younger patients. There isn't much trial information on this age group, so we can't, hand on heart, say there is a clear-cut benefit long term in the elderly."
"But we don't want to throw the baby out with the bath water," he added. "I think we need to conform to the guidelines, which recommend that everyone should have an antiplatelet drug after an MI or stroke to reduce recurrent ischemic events, even the elderly. However, we need to do more to prevent bleeding with these drugs."
He acknowledged that there is a question about how long aspirin should be given. "We could perhaps do a trial where aspirin is withdrawn after a year or so in those patients without active cardiovascular disease. That is a valid research question."
Oxford Vascular Study
For the study, the researchers analyzed bleeding events in 3166 patients (50% age 75 years or older) treated with antiplatelet drugs (mainly aspirin based) after a first transient ischemic attack, ischemic stroke, or MI in the Oxford Vascular Study, a population-based study of the incidence and outcome of all acute vascular events in a population of 92,728 individuals, irrespective of age, registered at nine general practices in Oxfordshire, United Kingdom.
Patients were started on antiplatelets from 2002 to 2012, with follow-up until 2013. Around 30% were receiving some sort of gastric protection (PPI or histamine-2 antagonist).
Results showed that 405 patients had first bleeding events (218 gastrointestinal, 45 intracranial, 142 other) during 13,509 patient-years of follow-up. The average annual risk for bleeding was 3.36%; the risk for major bleeding was 1.46%.
Risk of nonmajor bleeding was unrelated to age, but the annual risk for major bleeding increased steeply above age 70 years, reaching 4.1% at age 85 years or older, with similar patterns for both life-threatening and fatal bleeding. This finding reflects high risks of upper GI and intracranial bleeding at older ages.
The hazard ratio (HR) for major bleeding in patients aged 75 years or older was 3.10, and the HR for fatal bleeding in this age group was 5.53 (compared to those younger than age 75 years).
The same was true of major upper GI bleeds (=75 years: HR, 4.13), and the risk was particularly increased for disabling or fatal GI bleeding (HR, 10.26) in this age group.
There were 489 nonfatal and 208 fatal ischemic vascular events during follow-up. The ratio of major bleeds to ischemic events increased sharply with age: 0.20 in patients younger than 75 years, 0.32 for those aged 75 to 84 years, and 0.46 for those aged 85 years or older.
The authors further point out that the risk for major bleeds estimated to be attributable to antiplatelet treatment was similar to the risk for ischemic events estimated to have been prevented in the group 85 years of age or older.
They report that in the 75-years-or-older group, major upper GI bleeds were mostly disabling or fatal (45 of 73 patients [62%]) and were more likely to be disabling and fatal than recurrent ischemic strokes (101 of 213 patients [45%]).
Disabling or fatal GI bleeds also outnumbered disabling or fatal intracerebral hemorrhage (45 vs 18) in this age group.
The researchers tried to assess the benefit that routine PPI treatment may have in this population by using data from a meta-analysis of randomized trials of PPIs vs placebo in patients taking antiplatelet drugs (predominantly aspirin), in which PPI use reduced upper gastrointestinal bleeding by 74%.
Using this estimate, they estimate that the number need to treat (NNT) with PPIs to prevent one major upper GI bleed at 5 years' follow-up is 80 for patients younger than 65 years, 75 for patients aged 65 to 74 years, 23 for patients aged 75 to 84 years, and 21 for patients aged 85 years or older.
The NNT with PPIs to prevent one disabling or fatal upper GI bleed at 5 years' follow-up also decreased with age, from 338 for patients younger than 65 years to 25 for patients aged 85 years or older.
In a comment accompanying the Lancet paper, Hans-Christoph Diener, MD, University Duisburg-Essen, Essen, Germany, says that the benefit–risk association of long-term antiplatelet therapy should be regularly evaluated in patients older than 75 years, as is done with oral anticoagulants in patients with atrial fibrillation.
He agrees with the researchers that the study supports the use of PPIs in patients aged 75 years or older receiving antiplatelet therapy.
Dr Diener points out that PPIs are underused in patients receiving antiplatelet therapy, perhaps because the consequences of upper GI bleeds were underestimated in elderly patients who were treated with aspirin.
He also notes that a study published last year suggested a correlation between PPI use and dementia "was reported widely in the media and created a lot of confusion and angst" but was small and underpowered.
Professor Rothwell noted that PPIs prevent only GI bleeding, and there would still be other bleeds, the most serious of which are intracerebral hemorrhage (ICH). "This is an issue that needs to be considered, and we need to focus on patients' blood pressure to reduce the ICH risk. But on the whole non-GI bleeding is less preventable.
"Our recommendation is that all patients 75 years of age or over discharged after a stroke or MI on aspirin should be co-prescribed a PPI, and this should be continued long-term."
At present, patients are often prescribed aspirin at discharge, with hospital physicians thinking the family doctor may follow up with a PPI, but that doesn't necessarily happen, he said.
"Also, while we don't feel that all patients over 75 currently taking antiplatelet agents should be recalled at this point in order to start a PPI, I do think it is reasonable that when they do go to see their family doctor they are advised to start a PPI," he said. "Our data show that while the risk of bleeding is highest in the first couple of years it is still significantly elevated over the longer term."
The researchers estimate about half the population older than age 75 is taking aspirin (60% of the US population vs 40% of the UK population in this age group). The current study included patients taking aspirin for secondary prevention after having had a stroke, transient ischemic attack, or MI. Professor Rothwell estimates that these patients make up about two thirds of those taking aspirin, with the other third taking aspirin for primary prevention of cardiovascular disease.
"I think our results can be applied to the secondary prevention population immediately, but it is not clear whether the primary prevention population taking aspirin also need to be on a PPI, as they do seem to have a lower risk of bleeding."
The study was funded by the Wellcome Trust, Wolfson Foundation, British Heart Foundation, Dunhill Medical Trust, National Institute of Health Research (NIHR), and the NIHR Oxford Biomedical Research Centre. Dr Rothwell has received personal fees from Bayer outside the submitted work. Disclosures for Dr Diener are available with the paper. The other authors have disclosed no relevant financial relationships.