Whether and when to prolong dual antiplatelet therapy beyond 1 year after an acute coronary syndrome remains controversial.1 Continuation of dual antiplatelet therapy beyond 1 year reduces the risk of myocardial infarction and stent thrombosis, with the tradeoff of increased bleeding.2 Like all effective therapies, however, the potential benefits depend on the ability to adhere to the therapy.
Patients discontinue treatment for a variety of reasons. These include the perception that the treatment is not working, the expense, adverse effects, and simply forgetting to take the medication. Some adverse effects are intolerable (such as persistent coughing while receiving angiotensin-converting enzyme inhibitors) or are so dangerous (such as angioedema) that the risk of continuing treatment outweighs the potential benefit. Other side effects are self-limited, minor, or even favorable (eg, treating patients with hypertension with phosphodiesterase inhibitors and discovering that this type of treatment improves erectile dysfunction, or discovering that treatment with minoxidil has the side effect of hair growth, which is useful to treat male pattern baldness). But for unpleasant or even mildly harmful adverse effects, balancing the potential benefits of therapy against these adverse effects is a clinical challenge that is strongly influenced by patient perceptions and preferences. Further complicating the situation is that adverse effects are commonly ascribed to treatments even if symptoms are actually likely not due to the treatment, as is the case with muscle symptoms with statins.3,4
Observational cohort studies show that a complex set of medical, psychosocial, economic, and behavioral issues result in only about half of patients continuing to take their evidence-based medications a year after starting them.5 Even for a serious disease such as diabetes, it has been shown that 15% of patients never even fill their first prescription.6
In this issue of JAMA Cardiology, Bonaca and colleagues7 report the details of the reasons and timing of discontinuation of treatment with ticagrelor or placebo in the PEGASUS-TIMI 54 (Prevention of Cardiovascular Events in Patients With Prior Heart Attack Using Ticagrelor Compared to Placebo on a Background of Aspirin–Thrombolysis in Myocardial Infarction 54) trial. In this randomized clinical trial, patients who had a myocardial infarction between 1 and 3 years earlier were randomly assigned to ticagrelor or placebo for a median treatment period of 33 months. Nearly one-third of patients discontinued treatment with ticagrelor during that time period; perhaps even more interesting was that 21% of patients discontinued treatment with placebo.
The placebo comparator allows one to gain insight into the reasons for discontinuation specific to ticagrelor. Discontinuation rates were much higher in the PEGASUS-TIMI 54 trial than in the PLATO (Platelet Inhibition and Patient Outcomes) trial,8 presumably because minor symptoms are better tolerated and may be overlooked in the early phase of an acute coronary syndrome. Bleeding and dyspnea were each about 5 times more common as a reason for discontinuing treatment with ticagrelor than with placebo, with absolute discontinuation rates about 5% higher for patients who were receiving ticagrelor than for patients who were receiving placebo owing to each of these 2 adverse effects. The majority of the excess in discontinuation rates with regard to treatment with ticagrelor vs placebo was related to these 2 adverse effects. Dyspnea occurred early, with a median time to discontinuation due to dyspnea of about 10 days. Dyspnea was mild or moderate in severity in the majority of cases. This well-described adverse effect of ticagrelor does not manifest as abnormal pulmonary function test results and is quickly reversible after discontinuation. It does, however, require a thorough conversation because the patient and other health care professionals must be aware that the development of dyspnea ought not lead to anxiety, cessation of therapy without initiation of another P2Y12 inhibitor, or an expensive evaluation of patients who develop such symptoms in the 1 to 2 weeks after initiation of therapy.
Bleeding, the other leading cause of discontinuation of ticagrelor, was considered to be “nonmajor” in 86% of cases in which it was the reason for discontinuation. But bleeding was classified as minor with the decrease in hemoglobin level of less than 5 g/dL (to convert to grams per liter, multiply by 10.0), a definition that may not be minor from a patient’s perspective, especially since it was troublesome enough to lead the patient to stop a potentially lifesaving therapy. And type 1 bleeding (defined by the Bleeding Academic Research Consortium as “minor” bleeding that does not cause the patient to seek an unscheduled study or treatment) has been shown to be associated with a worse quality of life, even after adjusting for other clinical factors.9
Bonaca and colleagues7 performed a landmark analysis by evaluating discontinuation rates after the first year of treatment, after which time the rates of discontinuation were low and comparable between the patients who were receiving ticagrelor and the patients who were receiving placebo. This may be particularly helpful for clinical decision making because the results of the PEGASUS-TIMI 54 trial will most likely lead to the continuation of treatment with ticagrelor beyond a year for patients in whom ticagrelor was administered throughout the first year following the infarction. (The reinitiation of P2Y12 receptor antagonists more than a year after a myocardial infarction for patients in whom they had been discontinued is less likely to occur, in our estimation). The modestly lower discontinuation rate (and lower bleeding rate) for the lower (60 mg twice daily) dose of ticagrelor supports the selection of that dose if ticagrelor is used long term.
Bonaca and colleagues7 also report an “on-treatment” analysis to gain further insight into the potential effect of differential discontinuation on the estimation of the treatment effect. They suggest that intention-to-treat analyses may underestimate the true treatment effect that occurs among patients who continue treatment. While a sensitivity analysis showing a consistent or larger effect is reassuring, on-treatment analyses are confounded and may be misleading because the control group is no longer an ideal control group owing to differential dropout. The propensity analysis is an unsatisfactory, or at least incomplete, approach to addressing this confounding because no method of adjustment can account for the unmeasured confounders that exist.
How relevant is studying a clinical trial population to understanding adherence in unselected populations? Patients who volunteer to participate in trials are more adherent to study medications than unselected patients representing the general population. For example, patients enrolled in the STABILITY (Stabilization of Atherosclerotic Plaque by Initiation of Darapladib Therapy) trial (which evaluated darapladib on top of optimal medical therapy) were encouraged to continue statin therapy, and 96% did so for 3.7 years.10 In contrast, data from registries indicate that half of unselected patients stop taking statins by 1 year. A striking example of the discrepancy between adherence in clinical trials and adherence in clinical experience regards muscle aches believed related to statins, which occur no more commonly with statins than with placebo in large randomized trials3,11 but can lead to cessation of therapy in 15% to 20% of patients in practice.12 A rigorous placebo-controlled randomized clinical trial4 focusing on muscle symptoms showed that a high dose of atorvastatin doubled the incidence of reproducible muscle complaints, from 4.6% to 9.4% of participants, compared with placebo. Recent studies showing that most patients tolerate a rechallenge with statins13 provide further evidence that, most of the time, muscle aches experienced by patients receiving statins are not due to the statins.
The TRANSLATE-ACS (Treatment With Adenosine Diphosphate Receptor Inhibitors: Longitudinal Assessment of Treatment Patterns and Events After Acute Coronary Syndrome) study14 enrolled consenting US patients being discharged from the hospital after acute coronary syndrome to assess the use of and the persistence with secondary prevention treatment, and found that 92% of patients continued taking P2Y12 receptor antagonists at 6 months. Yet studies of unselected populations show that 14% of patients who had an acute myocardial infarction that was treated with coronary stents had stopped taking clopidogrel within 30 days after discharge from a group of US hospitals,15 and only half of the patients in a German cohort continued taking clopidogrel at 1 year.16 Therefore, to best understand adherence to medications and the adverse effects from them, a combination of placebo-controlled and registry data from an unselected cohort of patients is needed. The high frequency with which ticagrelor was discontinued in the PEGASUS-TIMI 54 trial may have been even higher if ticagrelor had been administered to patients who were similar to those found in routine clinical practice, emphasizing the importance of taking steps to promote adherence.
Thus, the development of systems to improve persistent use of secondary prevention therapies ought to be a high priority. Patient education and engagement with behavioral tools will be important when attempting to maximize adherence, especially to treatments that do not make patients feel better while they are receiving them, such as antiplatelet therapy. Improving adherence requires the ability to reliably measure it. Tools to organize medications, remind patients, and educate health care professionals (including pharmacists) and patients are proven ways to improve adherence17 and have been specifically shown to be effective after acute coronary syndromes.18 Knowledge of the expected adverse effects of treatment and of how to manage them is important to developing optimal educational approaches.
In summary, Bonaca and colleagues7 have shed light on the types and frequency of adverse effects leading to the discontinuation of treatment with ticagrelor (and placebo) in the years after an acute coronary syndrome. This information should be particularly helpful for clinicians as they apply the results of the PEGASUS-TIMI 54 trial for the benefit of their patients.
Corresponding Author: Christopher B. Granger, MD, Duke Clinical Research Institute, 2400 Pratt St, Room 7035, Durham, NC 27705 (email@example.com).
Published Online: June 15, 2016. doi:10.1001/jamacardio.2016.1018
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Granger reports receiving grants and personal fees from AstraZeneca, Bristol-Myers Squibb, Daiichi Sankyo, and Eli Lilly during the conduct of the study; grants from Armetheon, the US Food and Drug Administration, and the Medtronic Foundation outside the submitted work; grants and personal fees from Bayer, Boehringer Ingelheim, GlaxoSmithKline, Janssen Pharmaceuticals, The Medicines Company, Novartis, Pfizer, Sanofi, and Takeda outside the submitted work; personal fees from Gilead, Hoffmann–La Roche, Medtronic Inc, and the National Institutes of Health outside the submitted work. Dr Berger reports that, in the last 3 years, he has been the principal investigator of studies for which Northwell Health has received funding from AstraZeneca, The Medicines Company, Bristol-Myers Squibb, Sanofi, and Lilly/Daiichi Sankyo (all for >$10 000) outside the submitted work. No other disclosures were reported.
Christopher B. Granger, Peter B. Berger. Understanding the Adverse Effects of Ticagrelor in Practice. JAMA Cardiol. 2016;1(4):381–383. doi:10.1001/jamacardio.2016.1018