Aspirin Fails to Prevent CV Events in People at ‘Moderate’ Risk of Disease: ARRIVE

MUNICH, Germany—An age-old question—whether an aspirin a day truly keeps cardiovascular disease at bay in lower-risk individuals—remains unanswered, despite the efforts of a 12,000-patient randomized controlled trial that set out to settle the issue.

J. Michael Gaziano, MD (Brigham and Women’s Hospital, Boston, MA), presented the results of the ARRIVE trial at the European Society of Cardiology (ESC) Congress 2018; they were published simultaneously in the Lancet.

“While no overall reduction was observed in the primary composite endpoint of CV events, results from ARRIVE are generally consistent with many other studies that tended to demonstrate aspirin’s ability to lower the risk of first nonfatal MI without affecting risk of total stroke,” said Gaziano.

The role of aspirin in the secondary prevention of coronary disease and stroke, as well as in the setting of acute MI and stroke, is well established. But as Gaziano reminded his audience, the use of aspirin for primary prevention has far fewer trials to support its use, with most being done among patients at low risk of CVD. A recent large meta-analysis of aspirin in primary prevention suggested that a high risk of bleeding may offset the cardiovascular benefits in lower-risk individuals, and prompted the US Preventive Services Task Force to tighten up its recommendations on aspirin, saying it should be restricted to people at high risk of events.

ARRIVE therefore set out to establish the safety and efficacy of a daily dose of 100 mg, enteric-coated aspirin compared with placebo among 12,546 patients whose 10-year risk of a first acute cardiovascular event fell into the “moderate” range—between 10 to 20%. Patients at high risk for gastrointestinal bleeds or with diabetes were excluded from the trial.

Arriving at Aspirin

ARRIVE was conducted between July 2007 and November 2016 in seven countries and at more than 500 study sites, with follow-up conducted by primary care physicians during annual patient visits, and if those weren’t available then via phone calls and medical record checks.

At a median of 60 months, the observed cardiovascular event rate was “considerably less than anticipated,” Gaziano said: 9% as opposed to the expected 17%. An event rate of less than 10% corresponds to a lower-risk group—one for whom the role of aspirin has been better studied, he noted.

For the primary endpoint—time to first occurrence of cardiovascular death, MI, unstable angina, stroke, and transient ischemic attack (TIA)—there were no significant differences between the groups assigned to aspirin versus placebo (4.29% vs 4.48%; HR 0.96; 95% CI 0.81-1.13). Likewise, there were no differences in the secondary endpoints, made up of the individual components of the primary endpoint alone or in different combinations.

Overall rates of gastrointestinal bleeds were low, less than 1%, but twice as high in the aspirin group than in the placebo group, a statistically significant difference. Serious adverse events were relatively common (bleeding events as well as nonbleeding events such as osteoarthritis, cardiac events, and cancers), occurring in approximately one in five patients. Approximately 2.5% of patients in both groups died during the study. No significant effects were seen on short-term cancer, Graziano noted, adding that “the length of follow-up was insufficient to assess longer-term outcomes.”

In his presentation at ESC 2018, Gaziano zoomed in on a prespecified per-protocol analysis. Here again, the primary endpoint was no different between aspirin-treated versus placebo-treated patients; however, among patients who actually took aspirin compared to those who did not, aspirin was associated with a 47% relative decrease in the risk of first MI (HR 0.53; 95% CI 0.36-0.79) and a 45% relative decrease in the risk of first nonfatal MI (HR 0.55; 95% CI 0.36-0.84).

Asked by TCTMD why the intention to treat and per protocol analyses were so divergent, Gaziano pointed out that studying ASA in the 21st century is different than studying an investigational drug, given that aspirin is cheap and widely available.

“Our compliance rate overall in the per protocol analysis was about 60%, which is comparable to some of the other studies, but dropout was of a different flavor in this study than in previous studies,” he explained. “Not only were there disconsolations, which you see over time, which is random effect, but there is also selective drop-in. As we saw early in this study, people were developing TIAs and unstable angina, and even stable angina and chest pain—so developing risk. And what we saw was that individuals were being placed on aspirin either by themselves or by their physicians.”

This explains the substantial difference between the intention-to-treat and per protocol results, he continued, noting: “If you look at survival curves in the first 3 years, the curves begin to separate then they drift back towards each other.”

Investigators also saw signals of greater benefit among younger patients, and among patients at the lowest baseline risk, “again suggesting that in those individuals who were not being selectively treated with aspirin as conditions developed, there was the [beneficial] effects that we’ve seen in other previous trials,” Gaziano stressed. 

Also evident from ARRIVE is that predicting 10-year risk of cardiovascular events is not as simple as once seemed—a point made in an accompanying editorial, as well as by Gaziano himself to the press.

“Risk is not a static concept any more,” he said. “You start a study with an estimated risk of 17% and you end up with an observed risk of less than 9%: I think we should be proud that we are doing quite a bit [of good], but it does involve this complex risk-benefit assessment. I do believe that there are some patients whose risk calculus is sufficient to warrant aspirin as part of the armamentarium.”

Dose and Decisions

In the accompanying editorial, Davide Capodanno, MD, PhD (Policlinico Vittorio Emanuele, Catania, Italy), and Dominick J. Angiolillo, MD, PhD (University of Florida College of Medicine, Jacksonville, FL), point out that optimal aspirin dosing has been debated and is especially relevant in ARRIVE where body weights ranged from 34 kg to 177 kg. Mean body mass index in the trial was 28, Gaziano noted.

“Whether the same neutral results would be replicated by tailoring the dose of aspirin according to body weight remains a matter of interest,” Davide Capodanno and Dominick J Angiolillo write.

Indeed, further analyses of ARRIVE by body weight are planned, Gaziano confirmed. Another ongoing US trial—ADAPTABLE—is comparing different doses of aspirin, he continued, but for now, “the best that we can do in looking at dose these days is the large meta-analyses that are dominated by the secondary prevention trials that have enough data to come up with meaningful subgroups that can address that in an exploratory fashion.”

Applying to Practice

The overall negative results of the trial will likely continue to give pause for a drug that has been a mainstay of cardiovascular disease prevention for the past four decades, the editorialists conclude. “The consistent trend in negative results from trials of aspirin in primary prevention, particularly in patients without diabetes, suggests that new avenues of research are needed for the prevention of cardiovascular events,” Capodanno and Angiolillo write.

Experts discussing the results in a morning press conference, however, were somewhat more sanguine.

Stephan Achenbach, MD (University of Erlangen, Germany), pointed out that among cardiologists and preventionists there has been a “general atmosphere” of caution surrounding aspirin—“that it’s been used too liberally: pop a pill every day.” More recently, he said, “there’s been a swing towards being more careful [with aspirin use].” ARRIVE, he continued, may encourage physicians to tailor aspirin decisions to individual patients “and maybe we will swing back to being more positive about aspirin in primary prevention, as long as it is done right.”

Speaking with TCTMD, Achenbach said that further analyses of subgroups in ARRIVE will be complicated by the low event rates in the study paired with the low risk of the patients themselves. Still, he said, “this trial gives us data that again sound negative, but if you look at the analyses that are also available, for example the on-treatment analysis, it shows that there are some signals that heart attacks can be prevented, if you make individualized decisions. That’s something that gives us a little bit of hope that we will be able to make individual treatment regimens for our patients. It’s like lederhosen: they don’t fit everybody and in the same way aspirin is not for everyone. You have to select the right drug for the right person.”

Age and bleeding risk two key factors Achenbach said he will take into account when deciding whether or not to use aspirin. “Age and bleeding risk go along very closely together so . . . if you have a young patient with a high risk, that’s a patient where I would be more inclined to be using aspirin,” he explained. “Similar, again, to lederhosen, they suit the young people better than the old ones.”

ASCEND, another large study addressing aspirin in primary prevention, this one among patients with diabetes but no cardiovascular disease, is also coming out at ESC 2018, later today.