Long-Term Use of Antiplatelet Therapy in Real-World Patients with Acute Myocardial Infarction: Insights from the PIPER Study

The aim of this study was to assess long-term drug adherence and prognosis in real-world patients discharged on dual-antiplatelet therapy (DAPT) after acute myocardial infarction (AMI). A retrospective cohort analysis using administrative databases kept by eight local health units was performed. DAPT exposure (defined as ≥ 2 prescriptions), adherence, and the occurrence of major adverse events (MACE) were analyzed during a 36-month follow-up. The analysis included 11,101 patients who were discharged alive with a primary diagnosis of AMI. Of these, 5,919 patients (53.31%) were discharged on DAPT without a diagnosis of cancer or anemia, without transient DAPT discontinuation, and represented the study population. DAPT discontinuation occurred in 2,200 patients (37.2%) and in 1,995 (33.7%) after the first 6 and 12 months, respectively, whereas 423 patients (7.1%) were still on DAPT after 36 months. Patients who maintained DAPT up to 12 months had a significantly lower overall mortality, compared with patients who discontinued DAPT after 6 months. Exposure to DAPT at 3 years was associated with reduced all-cause mortality (hazard ratio [HR]: 0.067, 95% confidence interval [CI]: 0.027–0.162, p  < 0.001) and reduced recurrent AMI (HR: 0.02, 95% CI: 0.003–0.173, p  < 0.001). In conclusion, this study shows that prolonged DAPT over 12 months is maintained in a relevant number of patients after AMI. However, adherence to antiplatelet therapy in first 12 months after AMI is still unsatisfactory and efforts to enhance patients' compliance are warranted. Exposure to prolonged DAPT at 3 years seems to be associated with a significant reduction in all-cause mortality and AMI.

PEGASUS-TIMI 54 5 trials demonstrated that prolonged DAPT beyond 12 months can significantly reduce the incidence of major adverse cardiac events (MACE). In this setting, intensified antiplatelet therapy on top of aspirin has been shown to be an effective therapeutic strategy to prevent recurrent ischemic events. In selected patients at high thrombotic risk, prolonged DAPT beyond 12 months might be considered. [1][2][3] The need for prolonged DAPT might raise compliance issues, as long-term adherence to prescribed therapy after ACS is still challenging in real-world patients. 6 Previous reports have shown that more than 10% of patients prematurely discontinue antiplatelet therapy within 30 days after stent implantation. Clinical trials reported a rate of premature discontinuation as high as 30% after 12 months. 7 Premature antiplatelet therapy discontinuation raises safety concerns, as it is associated with a higher incidence of MACE. 6,8 Nonadherent patients are at a substantially higher risk of death. 9 Patients with ACS who discontinue all of their medications are more than three times as likely to die as those who remain adherent. Of note, the rate and clinical consequences of antiplatelet discontinuation beyond 12 months after ACS in real-world patients are still lacking.
The aim of our study is to assess long-term exposure and adherence to DAPT in patients with ACS in an Italian population. The occurrence of MACE in relation to maintenance of single-or dual-antiplatelet therapy has been investigated, as well.

Data Sources
The PIPER (Platelet-aggregation Inhibition: Persistence with treatment and cardiovascular Events in Real-world Study) is a retrospective, observational study, based on administrative databases of eight Italian Local Health Units (LHUs), in Lombardy, Tuscany, Puglia, and Campania, which includes approximately 5.6 million health-assisted individuals.
In particular, the following databases were used as data sources: the Health-Assisted Subjects' Database, containing patients' demographic data; Outpatients and Inpatients pharmaceutical drugs, providing information for each medication prescription, such as ATC (Anatomical-Therapeutic-Chemical) code of the drug purchased, number of packages, number of units per package, the dosages, unit cost per package, and prescription date; and Hospital Discharge Database, which includes all hospitalization data with discharge diagnosis codes classified according to the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). The patient code in each database permitted electronic linkage with all the other databases. No identifiers related to patients were provided to the researchers. Informed consent was not required for using encrypted retrospective information. This study was notified to the local ethics committee in each participating LHU according to the Italian law regarding the conduct of observational analysis, and the LHU Ethics Committees approved the study. 10

Study Populations
This was a retrospective cohort study that included all consecutive beneficiaries of each LHUs hospitalized and discharged alive with a primary diagnosis of acute myocardial infarction (AMI; ICD-9-CM: 410.xx) between January 1, 2010, and December 31, 2011 (enrollment period). The date of AMI discharge was identified as the index date, which represented the enrollment day of each individual patient, who was then followed up for 3 years after the index date (follow-up period).
Patients were classified according to exposure to antiplatelet therapy during the follow-up period. A patient was defined as treated with antiplatelet therapy (ATC code: B01AC), if at least two prescriptions of these drugs during the follow-up period were found. In particular, all enrolled patients were stratified into four main categories: (1) monotherapy with acetylsalicylic acid (ASA); (2) DAPT (ASA þ P2Y 12 inhibitors); (3) other drugs (only P2Y 12 inhibitors); and (iv) no antiplatelet therapy.
Patients with concomitant diagnosis at discharge of anemia or cancer were excluded, as these comorbidities might influence prescription and/or duration of DAPT.
Hospitalizations for AMI and all-cause mortality occurring during the 36-month follow-up period from discharge were considered. Multivariable analyses were conducted to check for possible confounders using a proportional hazards Cox regression model: demographic characteristics; other guidelinesrecommended post-AMI therapies, including beta-blockers, ACE inhibitors/AAIIs, statin, and antiplatelet therapy; previous hospitalizations for cardiovascular disorders; and diabetes.
A patient was defined as antidiabetic, antihypertensive, or cardiac therapy if at least two prescriptions of antidiabetic agents, of antihypertensives, of cardiac drugs, respectively, were found at follow-up. A patient was defined as treated with statins, beta-blocking agents, or ACE inhibitors/AAIIs if at least two prescriptions of these drugs were detected during the observation period. Patients who were transferred to another LHU during the observational period were excluded from analysis.
A Pegasus score was calculated for each patient by the sum of the characteristics which constituted the inclusion criteria of the Pegasus trial. 5 Cox regression analysis has been used for the risk assessment. The impact of different variables on cardiovascular mortality and recurrent AMI was analyzed using the hazard ratios (HR) with 95% confidence interval (CI); a p-value less than 0.05 was considered statistically significant. Analyses were performed with Stata software version 12.1 (Stata Corp LP, College Station, Texas, United States).

Results
A total of 11,101 were discharged alive with a primary diagnosis of AMI. Of these, 5,919 patients (53.3%) were discharged on DAPT, without a diagnosis of cancer or anemia at discharge, without transient DAPT discontinuation, and represented our study population (►Fig. 1).
The demographic and baseline clinical characteristics of included patients are shown in ►Table 1. Among the enrolled patients, 55.6% presented hypertension at the baseline, followed by 43.9% with heart disease, 26.0% with diabetes, and 10.0% with previous hospitalization for cardiovascular disorders. Statins, ACE inhibitors/AAIIs, and beta-blocking agents were used in 79.4, 74.6, and 71.5%, respectively. PCI at the index hospitalization was performed in 52.9% of patients. ►Table 2 describes the type of DAPT.
Baseline clinical and demographic characteristics of patients in relation to exposure and type of antiplatelet therapy are summarized in ►Table 3. Compared with patients who were still on DAPT at 6 months, patients who were on DAPT at 12 months were younger (mean age: 62.5 AE 16.0 vs. 65.5 AE 17.0, p < 0.001), less frequently required blood transfusion (3.9 vs. 6.9%, p < 0.001), and had less frequently chronic kidney disease (1.1 vs. 2.2%, p < 0.01). Patients who were still on DAPT at 3 years had more frequently a previous cardiovascular hospitalization (8.8 vs. 14.4%, p < 0.001) and were more frequently diabetic (38.5 vs. 25.5%, p < 0.001), compared with patients who discontinued after 12 months.

Discussion
The main findings of this study are the following: adherence to DAPT within the first 12 months after AMI is still suboptimal; notably, a significant proportion of patients in Italy maintain DAPT up to 36 months after AMI; cessation of DAPT within the first 12 months is associated with increased overall mortality; prolonged DAPT beyond 12 months seems to be associated with a significant reduction in all-cause mortality and AMI.
The benefit of DAPT after an ACS was well demonstrated in the CURE, 11 COMMIT/CCS-28, 12 and CLARITY-TIMI 28 13 trials. The association of aspirin and clopidogrel reduced the 1-year incidence of cardiovascular events by approximately 20% compared with aspirin alone. More potent and consistent  16,17 which demonstrated a sustained increased risk of thrombotic complications, including stent thrombosis and spontaneous cardiovascular events, beyond 6 months. Notably, the greatest absolute reductions in cardiovascular events with DAPT are seen in the first 3 months and since these studies, advances in drug-eluting stent technology have led to a substantially reduced incidence of late (>30 days) and very late (>1 year) stent thrombosis. 10,18 Although it has been fully elucidated that premature discontinuation of antiplatelet therapy is associated with a poor prognosis, patients' compliance remains low. 6,7 Several reasons for noncompliance have been reported, such as excessive concerns about hemorrhagic complications or competing guidelines, complex treatment regimens, or lack of outcome expectancy. 19 As anemia can be considered a marker for high bleeding risk, thus contraindicating prolonged DAPT, patients with anemia at discharge were excluded from the present analysis. Nevertheless, in this study, one out of three patients discontinued DAPT after 6 months. In routine clinical practice, poor medication adherence is one of the main factors that reduce the effectiveness of chronic drug therapies. 20 In accordance with data observed in other studies, 6,21,22 in this study a considerable proportion of patients report low adherence to prescribed medications, particularly for non-monotherapy strategies.
Although many studies have evaluated the incidence or effect of antiplatelet therapy cessation on subsequent cardiovascular risk, most included selected cohorts of patients and were limited by lack of standardized definitions for therapy discontinuation that did not integrate the underlying context in which treatment was discontinued. 23,24 The results presented here suggest that there are still numerous barriers to adherence to practice guidelines; therefore, encouragement to continue long-term therapy with lifestyle modifications and pharmacological treatments is needed to improve outcomes in patients at risk of cardiovascular events. 25 On the other hand, this study demonstrated that the number of patients who maintain DAPT up to 3 years after   benefit at the cost of a significant bleeding risk. European 26 and North American 27 guidelines therefore do not recommend DAPT in patients with stable atherothrombotic disease but acknowledge that with careful consideration, combined antiplatelet therapy may be beneficial in some high-risk patients.
In this study, prolonged DAPT up to 3 years after AMI was associated with a significant reduction in overall mortality and AMI. Due to the retrospective nature of the study, this conclusion should be interpreted with caution and only as hypothesis generating. It should be highlighted that prolonged DAPT might identify patient at lower hemorrhagic risk, without noncardiac comorbidities which would have contraindicated long-term DAPT. To reduce a selection bias, patients with anemia of cancer at discharge were excluded, as well as patients who were not prescribed DAPT at discharge. An unexpected finding in the DAPT trial 4 was a borderlinesignificant increase in overall mortality rate (0.5% absolute increase) with 30 months of DAPTversus 12 months of DAPT in DES-treated patients, which was due to significantly increased deaths from non-cardiovascular causes (most commonly cancer), with no increase in cardiovascular deaths, and no significant increase in fatal bleeding.
It should be highlighted that the prescription of DAPT at discharge was relatively low. This might be partly explained by the enrollment period, which was from January 2010 to December 2011, when P2Y 12 inhibitors in medically managed patients were still underused. Moreover, the analysis included not only type I but also type II MI patients, for whom  the benefit of DAPT may be controversial. The low use of beta-blockers and ACE inhibitors was evident, as well. The low use of ACE inhibitors might be due in some cases to preserved ejection fraction.
The results of this study should be interpreted in light of some strengths and limitations. The strength of the presented data includes its basis on the large size unselected population; our patients may represent a more representative sample of "real world" than those included in many of the randomized controlled trials that have previously evaluated antiplatelet use for mortality reduction. Limitations include, first, the absence of relevant clinical information in the data setting. It is plausible that patients' risk profile might have conditioned DAPT duration. However, due to the nature of the study, which consists of a retrospective analysis from administrative databases, important clinical data are missing and confounding biases might have influenced the results. To overcome this limitation, we excluded patients with concomitant diagnosis at discharge of anemia or cancer, as these comorbidities might influence prescription and/or duration of DAPT. It cannot be ruled out if the reduced allcause mortality associated with prolonged exposure to DAPT is related to a lower risk profile of patients (in terms of comorbidities), which allowed prolonged DAPT, or the protective effect of DAPT. Notably, prolonged DAPT was associated with a reduced rate of recurrent AMI, for which the protective effect of DAPT might be more evident. Second, adherence was estimated using pharmacy data on filled prescriptions, but no information on actual medication consumption was available. Third, the reasons for nonadherence were not retrievable from the dataset.
Despite these limitations, this study indicates prolonged DAPT beyond 12 months is maintained in a relevant number of patients after AMI. Notably, long-term exposure and adherence with antiplatelet therapy in patients with AMI still appear to be unsatisfactory in an unselected Italian population. Further efforts to enhance patients' compliance to medical therapy are warranted.
Conflicts of Interest L.D.E., V.P., C.V., and S.B. report no conflicts of interest in this work. Manuscript development was supported by an unconditional grant from AstraZeneca. R.R. reports payment as an individual for consulting fee or honorarium from Daiichi Sankyo, Inc, Astra Zeneca, Bayer, and Pfizer.

Authors' Contributions
All authors contributed toward data analysis, drafting, and revising the manuscript and agree to be accountable for all aspects of the work.

Ethical Approval
This article does not contain any studies with animals performed by any of the authors.

Informed Consent
Informed consent is not required by the Local Health Unit (LHU) ethics committees for using encrypted retrospective information. No identifiers related to patients were provided to the researchers. This study was notified to the local ethics committee in each participating LHU according to the Italian law regarding the conduct of observational analysis, and the LHU Ethics Committees approved the study.