High- and Low-Dose Aspirin Equivalent in ASCVD: ADAPTABLE
The open-label study had a high crossover rate, throwing some doubt on the reliability of the results, say experts.
(UPDATED) Treating patients with atherosclerotic cardiovascular disease (ASCVD) with low-dose aspirin is just as effective as using the 325-mg dose for the prevention of major cardiovascular events, and both doses are associated with a very low rate of bleeding, according to findings from the open-label, pragmatic ADAPTABLE trial.
The results, presented as a late-breaking clinical trial during the American College of Cardiology 2021 Scientific Session and published simultaneously in the New England Journal of Medicine, suggest equivalence between the two doses and should help inform clinical practice in the United States, where there is some uncertainty about the right aspirin dose for secondary prevention, say investigators.
Still, there are some question marks around the reliability of these new findings, particularly given the high rate of crossover from the high- to low-dose aspirin groups.
“The results showed there was no difference between the 81-mg and 325-mg doses,” lead investigator W. Schuyler Jones, MD (Duke University Medical Center, Durham, NC), told TCTMD. “There was some dose-switching in the 325-mg group which does make it a little bit tough to figure out the absolute correct answer, but we think, in general, that this answers the question, especially for the majority of patients. For those on the 81-mg dose, they should continue with 81 mg of daily aspirin.”
For patients currently taking the 325-mg dose, it’s a more-nuanced decision. If they tolerate the higher dose, then they might do better continuing to take it, Jones said, especially since the overall rate of bleeding is very low. “In those situations, we’re telling patients to talk to their clinicians because there is likely a reason they tolerated it well over time and we actually think that dose-switching is probably not a good idea for patients.”
John McEvoy, MB BCh (National University of Ireland, Galway), who led the review of aspirin for the 2019 American College of Cardiology/American Heart Association (ACC/AHA) primary prevention guidelines, praised the study for its innovative pragmatic design, calling the effort an important step forward for US-based trials that utilize electronic health records (EHRs).
The results showed there was no difference between the 81-mg and 325-mg doses. W. Schuyler Jones
“As a result of this trial, I hope we see more patients move towards a low-dose aspirin strategy for the secondary prevention of cardiovascular disease,” he told TCTMD. “I don’t think based on this study, and based on prior studies, there’s strong enough evidence to justify the ongoing use of high-dose aspirin for these patients.”
However, like Jones, McEvoy said there are caveats to that interpretation, noting that one of the drawbacks of the open-label study was that patients were unblinded to treatment. During the course of the trial, which ran from 2016 to 2019, more than 40% of patients treated with 325 mg switched to the lower dose. Given the large number of patients dropping down to 81 mg, as well as the 7.1% who switched from the low dose to the higher dose, there is some uncertainty about the findings.
“Although it’s an innovative, important study that builds for the future, in some aspects the nature of how the trial was run, specifically how patients in the trial maintained their randomized allocation, was disappointing,” he said. “There was a really significant crossover from the high-dose aspirin to the low-dose aspirin. You could argue that on a per-protocol or on-treatment basis, there may have been some benefit or risk for the full dose that wasn’t revealed by this study. In some ways, they ended up comparing low-dose against low-dose for many patients.”
Colin Baigent, BM BCh (University of Oxford, England), who wrote an editorial accompanying the study, believes the high crossover rate makes it impossible to conclude there is no significant difference between the two aspirin doses. “The ADAPTABLE experience is a reminder of the importance of certain fundamentals when designing and conducting randomized trials,” he writes. However, he acknowledges that when planning a new trial, it can be difficult to know where problems might arise.
During the late-breaking trial session, Donald Lloyd-Jones, MD (Northwestern University Feinberg School of Medicine, Chicago, IL), one of the discussants, called ADAPTABLE a pioneering pragmatic trial, a type that researchers will be doing more and more in the years to come. While the open-label design is a limitation given the crossover from high- to low-dose aspirin, he said, “the most important legacy of this trial is going to be that you did it and that you showed us many of the promises and some of the pitfalls of these large, pragmatic designs.”
Equipoise at Time of Trial Design
The European Society of Cardiology currently recommends low-dose aspirin for patients with stable ASCVD, but the ACC/AHA clinical guidelines do not take a definitive stance on aspirin dosage in such patients. At the time ADAPTABLE was designed, data from the National Cardiovascular Data Registry showed that more than 60% of patients discharged after MI were treated with the 325-mg dose, Jones noted.
“There was definitely some uncertainty about the dose,” he said. “Over the course of the last 10 years, though, things have shifted more toward the 81-mg dose based on the European guidelines and other guidelines in the US. At the beginning of the study, though, we thought there was equipoise.”
The study was conducted across 40 US centers participating in the national patient-centered clinical research network (PCORnet) and patients were recruited through the EHR. Eligible patients accessed a web portal to give informed consent and were then assigned to treatment—either the 81-mg or 325-mg dose that they purchased over the counter. Follow-up visits occurred every 3 or 6 months and all clinical outcomes were ascertained remotely and without adjudication.
In total, 15,076 patients with ASCVD were enrolled and randomized to treatment with a high or low dose. The median age was 67.6 years, the majority of patients were men (68.7%), and 8.7% were Black, 3.2% were Hispanic, and 1.0% were Asian. At baseline, one-third of patients had a prior MI and 53.0% had a previous coronary revascularization procedure within 5 years of enrollment.
In some ways, they ended up comparing low-dose against low-dose for many patients. John McEvoy
After a median follow-up of 26 months, death from any cause, hospitalization for MI, or hospitalization for stroke occurred in 7.28% of patients treated with the 81-mg dose and 7.51% of patients who received the 325-mg dose (HR 1.02; 95% CI 0.91-1.14). Taken individually, there was no difference in the risk of any of the endpoints, nor was there any difference in the risk of hospitalization for TIA or PCI/CABG surgery.
Similarly, there was no difference in the primary safety endpoint. The rates of major bleeding requiring blood-product transfusion were 0.63% and 0.60% with the 81-mg and 325-mg doses, respectively (HR 1.18; 95% 0.79-1.77).
Speaking during the session, Jones said the reasons for aspirin discontinuation/switching was quite heterogenous, ranging from patient preference to changes initiated by a physician. In a sensitivity analysis that focused on patients who adhered to the randomized aspirin dose, the risk of death, MI, or stroke was higher among those who took the lower dose (HR 1.25; 95% CI 1.10-1.43), but the researchers urge caution when interpreting that analysis given the potential for bias.
Akshay Desai, MD (Brigham and Women’s Hospital, Boston, MA), another of the study discussants, noted that the overall rate of major bleeding in the trial was quite low, citing the OASIS-7 study as just one example where bleeding rates were higher than in ADAPTABLE. Noting that the bleeding endpoint was somewhat unconventional, he questioned what steps investigators took to ensure complete ascertainment of bleeding events in the trial.
Jones noted that little about ADAPTABLE was conventional, and said that while there was no clinical events committee to adjudicate events, they captured events from multiple data sources, including EHRs, claims data, and patient reports.
“It was done in a different way, but we think the collection of events was pretty systematic,” said Jones.
Crossover and Dose-Switching
At the time ADAPTABLE ran, several studies testing aspirin in primary prevention were published, among them ASPREE, ASCEND, and ARRIVE, and the consensus from those studies was that aspirin was largely unnecessary in the setting of primary prevention. While the patients in ADAPTABLE all had established ASCVD, it’s possible the results from those studies influenced adherence to medication dose in the trial.
“We definitely think that patients’ perspectives or preferences might have played into the dose-switching and adherence,” said Jones. “Since three of the primary-prevention studies were published during the course of our study, we had a lot of confusion, not just in patients, but also in clinicians who thought that aspirin wasn’t useful. We definitely want people coming away from ADAPTABLE thinking aspirin is still useful. We think the 81-mg dose is the most-useful dose in secondary prevention.
To TCTMD, McEvoy said the development of PCORnet, which includes 29 health networks, will allow researchers to conduct future pragmatic, EHR-embedded randomized trials like ADAPTABLE. Studies such as these are common in countries with more-connected healthcare systems, such as Sweden, but have not been possible in the fragmented US healthcare system. One of the next steps is to learn how best to improve adherence in open-label studies, he said.
“In many ways, the aspirin question asked here was set up for crossover because aspirin is a medication that has a lot of baggage,” he said. “There are many patients out there that know about the risk of bleeding with aspirin and have concerns about aspirin in terms of its known side effects. When you’re doing a trial comparing high- versus low-dose aspirin, and patients know that’s how the trial is set up, and you’re compelled to go through the risks and benefits as part of the consent process, we’re kind of priming patients that there’s risks to aspirin.”
Nonetheless, McEvoy congratulated the researchers for their efforts, and is optimistic that future pragmatic trials can address issues around adherence. What is learned from ADAPTABLE will be valuable for future large, pragmatic, EHR-embedded studies, he said. In his editorial, Baigent also called the trial a major achievement for showing that a low-cost pragmatic trial could be conducted in the US.
“The method can now be adapted and used more widely,” writes Baigent.
Going forward, Jones suspects it might be possible mix traditional clinical trial methodology with a pragmatic design. For example, patients might still be recruited, enrolled, and tracked via the EHR, but medications could be mailed to those randomized to treatment, which should aid in adherence to medication.
While the design does have limitations, the pragmatic study also has numerous strengths, said Jones. Through the EHR, they were able to identify more than 450,000 eligible patients who were asked to participate in the trial. Moreover, by relying on the EHR, and other patient-reported information, they were able to complete the randomized trial at a significantly lower cost compared with typical trials. He added that they used nine “patient-partners” in the trial, which were patients who reviewed the study protocol, protocol amendments, and all patient-facing materials. They also provided input on all matters related to the study.
“We definitely think it’s the beginning about how we do [pragmatic trials] in the US,” said Jones. “We’re excited about it.”
Jones WS, Mulder H, Wruck LM, et al. Comparative effectiveness of aspirin dosing in cardiovascular disease. N Engl J Med. 2021;Epub ahead of print.
Baigent C. Pragmatic trials—need for ADAPTABLE design. N Engl J Med. 2021;Epub ahead of print.
- Jones reports grants from the Patient-Centered Outcomes Research Institute, Boehringer Ingelheim, and Merck; grants/personal fees from Bayer and Janssen Pharmaceuticals; and personal fees from Bristol-Myers Squibb and Medscape.
- Baigent reports grants from Boehringer Ingelheim, the Medical Research Council, the British Heart Foundation, and the National Institute for Health Research.