Skip main navigation

Days Alive and Out of Hospital: Exploring a Patient-Centered, Pragmatic Outcome in a Clinical Trial of Patients With Acute Coronary Syndromes

Originally published Cardiovascular Quality and Outcomes. 2018;11:e004755



    Cardiovascular clinical trials have traditionally incorporated separate time-to-first-event analyses for their primary efficacy and safety comparisons, but this framework has a number of limitations, including limited patient-centeredness and a traditional requirement for central adjudication. Days alive and out of the hospital (DAOH) has the potential to provide additional insight.

    Methods and Results

    TRILOGY ACS (Targeted Platelet Inhibition to Clarify the Optimal Strategy to Medically Manage Acute Coronary Syndromes) was a randomized, multinational clinical trial that compared the effect of prasugrel versus clopidogrel in patients stabilized after non-ST segment elevation acute coronary syndrome treated without revascularization; the trial had a neutral result. DAOH was calculated for each patient using site-submitted adverse event reporting data. We described patterns of DAOH overall, and among younger adults (<75 years old), older adults (≥75 years old), and frail/prefrail patients over 12 months follow-up and used Poisson regression to compare DAOH for patients randomized to prasugrel versus clopidogrel. Of 9249 patients in the overall trial population, 500 (5.4%) died, and 2504 (27.1%) were hospitalized 4150 times over 12 months’ follow-up; the mean±SD DAOH was 317±86. The distribution of DAOH over 12 months was left-skewed, with median DAOH 363 days. Among younger adults, older adults, and frail/prefrail patients, mean DAOH were 323, 293, and 304 days, respectively. There were no differences in DAOH by treatment arm in the overall population (rate ratio, 1.00; 95% CI, 0.99–1.01) or any subgroup.


    These results support the feasibility of determining DAOH, a patient-centered outcome that can potentially overcome many of the disadvantages of the traditional time-to-composite-event framework in the clinical trial setting.

    Clinical Trial Registration

    URL: Unique identifier: NCT00699998.


    • The traditional time-to-composite–end point framework for analyzing clinical trials of antithrombotic agents has a number of limitations, including inability to account for bleeding and ischemic events in the same analysis, limited patient-centeredness, and a requirement for central adjudication.

    • Analyzing the effect of a treatment on days alive and out of the hospital, in addition to or instead of time to ischemic events, has the potential to overcome some of these limitations.


    • Days alive and out of the hospital can be feasibly calculated from site-submitted safety data and used to compare treatments in a randomized controlled trial of antithrombotic agents for patients with acute coronary syndromes.

    • In TRILOGY ACS (Targeted Platelet Inhibition to Clarify the Optimal Strategy to Medically Manage Acute Coronary Syndromes), a randomized controlled trial of prasugrel versus clopidogrel in patients with medically managed acute coronary syndrome, the days alive and out of hospital framework replicated the results of the traditional time-to-composite event framework, showing no difference between treatment arms.

    • Days alive and out of the hospital may represent a more patient-centered framework for analyzing clinical trial results that is useful in pragmatic clinical trials.


    Clinical trials of antithrombotic agents typically have a primary efficacy outcome of time to the composite of cardiovascular death, myocardial infarction (MI), or stroke, and a primary safety outcome of major bleeding by one of several criteria. These end points are ill-suited to guide patient and physician decision-making. Increasingly potent combinations of antithrombotic therapy increase bleeding events in concert with reductions in ischemic events,1–3 and interpretation of these opposing effects has resulted in ambiguity and disparity in guideline recommendations.4,5 More broadly, the clinical significance of composite end points is uncertain, and neither ischemic events nor major bleeding represent truly patient-centered outcomes because by lumping together all patients with a given diagnosis, they do not account for severity and fail to capture the varying health consequences faced by individual patients with these diagnoses.6 Furthermore, the central adjudication process usually required to identify ischemic and bleeding events increases the cost and complexity of clinical trial conduct.

    Several useful alternate approaches for analyzing clinical trial results have been proposed, but all require central event adjudication and make a priori assumptions about the relative severity of event types that may not be shared by all patients, physicians, and regulatory authorities.7,8 From the patient perspective, a common thread linking ischemic and bleeding events are death and days spent in the hospital. In instances where a hospitalization has downstream consequences, these consequences ultimately result in further hospitalization or death. Thus, days alive and out of the hospital (DAOH) represents an important patient-centered outcome that approximates time spent in good health. It accounts for multiple events over the course of a study period, weights death more heavily than hospitalization, and weights deaths occurring early after trial enrollment more heavily than those occurring later. Because it does not require event adjudication, it is also an attractive end point for pragmatic clinical trials.

    DAOH has been evaluated in a secondary analysis of a heart failure trial,9 but it has not been evaluated in the context of a trial enrolling patients with acute coronary syndromes (ACS), in which clinical end point events are less common and adverse events leading to hospitalization are more common compared with heart failure trials. In this analysis, we use data from TRILOGY ACS (Targeted Platelet Inhibition to Clarify the Optimal Strategy to Medically Manage Acute Coronary Syndromes), a recently completed randomized clinical trial, to demonstrate the use of DAOH as an ACS clinical trial end point.


    The TRILOGY ACS trial design and primary results have been previously published.10,11 Briefly, TRILOGY ACS was a randomized, double-blind, double-dummy trial that compared the safety and efficacy of up to 30 months of treatment with aspirin+prasugrel versus aspirin+clopidogrel in patients with non-ST segment elevation-ACS (NSTE-ACS) who did not undergo in-hospital revascularization. The trial found no difference between prasugrel and clopidogrel in the incidence of its primary efficacy outcome (time to first occurrence of the composite of cardiovascular death, MI, or stroke) or safety outcome (major bleeding not related to coronary artery bypass grafting by TIMI [Thrombolysis in Myocardial Infarction] criteria). Patients could undergo coronary angiography before enrollment in the trial but were only eligible for enrollment if they were selected for a final strategy of medical management without revascularization within 10 days of their qualifying ACS event. Enrolled patients received a loading dose of either 300 mg clopidogrel or 30 mg prasugrel, followed by daily blinded maintenance doses of either 75 mg clopidogrel or 10 mg prasugrel (5 mg for patients ≥75 years old or <60 kg body weight), which was continued for at least 6 months and up to 30 months. Institutional review boards at each participating center approved the TRILOGY ACS study, and all participants gave informed consent. The data, analytic methods, and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedure.

    Patient Population

    TRILOGY ACS enrolled 9326 patients in 52 countries. For our study, we excluded patients that died in hospital (n=25; 9 patients randomized to prasugrel and 16 randomized to clopidogrel), had no outpatient follow-up (randomization date=date of last follow-up or index hospitalization discharge date was later than the date of last follow-up, n=104), or were missing the date of their index ACS event (n=2), yielding 9249 patients in the overall population. We separately analyzed younger adults (patients <75 years old, analogous to the primary analysis population included in the main TRILOGY ACS manuscript, n=7192), older adults (≥75 years old, n=2057), and the frail/prefrail population (n=1361; Figure 1). Frailty was collected only for patients ≥65 years old and was defined by patient report of at least 3 phenotypic characteristics consistent with reduced vitality at baseline: unintentional weight loss, decreased grip strength, increased fatigue/lethargy or declining endurance, slow gait speed, and decline in typical physical activity level.12,13 Patients meeting 1 or 2 criteria were classified as prefrail. Because of the low number of patients enrolled in the study meeting criteria for frailty and prefrailty, frail and prefrail patients were grouped together into a single frail/prefrail population.

    Figure 1.

    Figure 1. Study flow.

    DAOH Definition

    DAOH was calculated separately for each patient enrolled in the trial using adverse event data submitted by sites, which captures hospitalizations and deaths, rather than centrally adjudicated clinical end point event data. First, total potential follow-up time was determined for each patient; total potential follow-up time was determined as number of days from the index event discharge date until the date of the final patient examination, if the patient completed the study, withdrew, or was lost to follow-up, or the end of study, if the patient died. Follow-up started at the time of discharge because TRILOGY ACS enrolled patients stabilized following NSTE-ACS, with median time from admission to randomization of 4 days. As a sensitivity analysis, we repeated our analyses with follow-up starting at the time of randomization. Patients were followed for a maximum of 30 months in TRILOGY ACS, but median follow-up was 17 months, and Poisson regression can give a biased estimate of difference between treatment groups if too many patients are censored before completion of follow-up.14 Because most patients completed 12 months follow-up, follow-up was truncated for all patients at a maximum of 12 months to limit issues with bias resulting from differential censoring. Twelve-month follow-up is also consistent with consensus guideline recommendations on duration of treatment with aspirin plus a P2Y12 inhibitor in patients with ACS.4,5,15,16 The total number of days spent in hospital was derived by checking hospitalization status on a daily basis beginning with the day after the index event discharge day. Hospitalization was defined as inpatient admission with length of stay ≥24 hours. If a patient died, the number of days from their death to the end of study follow-up (up to 365 days) were counted. DAOH was calculated by subtracting days in hospital and days from death until end of study follow-up from total potential follow-up time.

    Statistical Analysis

    For the overall population, <75-year-old population, older adult population, and frail population, histograms, and empirical density curves were created to show the distribution of DAOH for all enrolled patients over 12 months follow-up and separately for patients randomized to clopidogrel and prasugrel. In the overall population, we also calculated the median (25th, 75th percentile) number of DAOH for patients with and without ischemic and bleeding events. We also created figures for each population showing the proportion of patients alive and out of the hospital, hospitalized, and dead on a daily basis over the course of follow-up.

    In the overall population, we compared DAOH over 12 months follow-up for patients randomized to clopidogrel to patients randomized to prasugrel using Poisson regression with a robust variance estimate and an offset term to account for differential lengths of follow-up. This procedure yields a rate ratio and 95% CI, which can be thought of as DAOH in the prasugrel arm divided by DAOH in the clopidogrel arm. As clinical practice patterns differ globally, a second model was constructed to test the interaction of randomized treatment and geographic region on DAOH. Though the interaction was nonsignificant, we nevertheless report region-specific effect estimates and CIs because the study may be underpowered to detect an interaction. To identify differences in treatment effect by patient subgroup, we repeated the primary analysis in the younger adult population, the older adult population, and the frail/prefrail population.


    Characteristics of patients enrolled in TRILOGY ACS have been previously reported.10 In all populations evaluated, baseline patient characteristics were similar in the prasugrel and clopidogrel treatment arms.

    Deaths and Hospitalizations by Subgroup

    Of 9249 patients in the overall population, 500 (5.4%) died, and 2504 (27.1%) were hospitalized 4150 times over 12 months’ follow-up; 153 patients died without being hospitalized, 347 were hospitalized and then died. Six thousand five hundred and ninety-two patients (71.2%) had neither a death nor hospitalization. Among 2504 patients that were hospitalized, 1563 (62.4%) had 1 hospitalization, 573 (22.9%) had 2, and 368 (14.7%) had ≥3. The mean (SD) number of hospitalizations per patient was 0.4 (1.0). By comparison, there were 724 MI events, 90 stroke events, and 123 TIMI major or minor bleeding events through 12 months’ follow-up; these events accounted for 22.6% of all hospitalizations.

    In the younger adult (<75-year-old) population (n=7192), 276 (3.8%) patients died and 1731 (24.1%) were hospitalized; the mean (SD) number of hospitalizations per patient was 0.4 (0.9). Among older adult patients (n=2507), 224 (8.9%) died, and 773 (30.8%) were hospitalized, with a mean (SD) number of hospitalizations of 0.6 (1.1). Of 1361 frail/prefrail patients, 129 (9.5%) died, and 498 (36.6%) were hospitalized at least once, with a mean (SD) number of hospitalizations of 0.6 (1.2).

    DAOH by Subgroup

    In the overall population, mean±SD DAOH over 1-year follow-up was 317±86 days with a left-skewed distribution; median DAOH was 363 (25th, 75th percentiles: 328, 365; Table 1, Figure 2A). Among patients that had a TRILOGY ACS primary end point event (cardiovascular death, MI, or stroke) within the first 12 months of follow-up, median DAOH was 260 (25th, 75th percentiles: 118, 348) compared with 365 (333, 365) in patients without an end point event. Patients with any component of the primary composite ischemic end point or a bleeding end point had fewer DAOH than patients without (Table I in the Data Supplement).

    Table 1. DAOH by Subgroup

    Patient PopulationNDAOHDays DeadDays Hospitalized
    Mean (SD)Median (IQR)Mean (SD)Median (IQR)Mean (SD)Median (IQR)
    Overall9249317 (86)363 (306–365)12 (54)0 (0–0)4 (10)0 (0–0)
    Younger adult (<75-y old)7192323 (79)365 (328–365)8 (44)0 (0–0)3 (9)0 (0–0)
    Older adult (≥75-y old)2057293 (105)357 (242–365)25 (78)0 (0–0)6 (14)0 (0–2)
    Frail/prefrail1361304 (99)359 (268–365)20 (69)0 (0–0)6 (13)0 (0–6)

    DAOH indicates days alive and out of the hospital; and IQR, interquartile range.

    Figure 2.

    Figure 2. Distribution of days alive and out of the hospital. Histogram and overlaid empirical density distributions show the distribution of days alive and out of the hospital (DAOH) for (A) all enrolled patients (B) patients <75-y old, (C) patients ≥75-y old, and (D) frail/prefrail patients randomized to clopidogrel (blue) and prasugrel (red).

    In the younger adult (<75-year-old) population, mean DAOH was 323±79 days (median 365; 25th, 75th percentiles 306, 365); mean DAOH were 293±105 and 304±99 days in the older adult (≥75-year-old) and frail/prefrail populations, respectively (Table 1, Figure 2B–2D). The relative contributions of death and hospitalization to lost DAOH were different in each subgroup: For example, in the overall population the mean±SD number of days from death until end of study follow-up was 12±54, and the mean±SD number of days hospitalized was 4±10; in the older adult population, the mean±SD number of days from death until end of study follow-up was 25±78, and the mean±SD number of days hospitalized was 6±14.

    There were marked differences in mean DAOH by region (ranging from 298 in the Mediterranean basin to 340 in the Indian subcontinent), potentially attributable to differences in baseline patient characteristics as well as regional differences in patient management and patterns of hospitalization post-ACS (Table II in the Data Supplement).

    When evaluating the proportion of patients in a given state each day during follow-up, we found that patients died at a steady rate through follow-up but that most hospitalizations were clustered in the early part of follow-up. The net effect was that the proportion of patients alive and out of the hospital was highest immediately after the index admission and declined steadily over the course of follow-up (Figure 3A). This pattern was evident in the younger adult population, older adult population, and frail population, though older and frail patients were more likely to die and more likely to be hospitalized than patients in the overall population (Figure 3B–3D).

    Figure 3.

    Figure 3. Proportion of patients in various states (dead, hospitalized, or alive and out of hospital) on a daily basis through 12 mo of follow-up. Graphics represent 365 individual bar graphs showing the proportion of patients dead (blue), hospitalized (gray), or alive and out of the hospital (orange) on each day of follow-up among (A) all enrolled patients, (B) patients <75-y old, (C) patients ≥75-y old, and (D) frail/prefrail patients. Patients that die remain dead (and so the blue area steadily increases in size over time), but patients can cycle back and forth between hospitalized (gray) and alive and out of the hospital (orange).

    DAOH by Treatment Arm

    In the overall population, patients randomized to clopidogrel had a mean 317±86 DAOH over 12 months follow-up compared with 316±87 for patients randomized to prasugrel. There was no difference in the proportion of potential follow-up time spent alive and out of the hospital in patients randomized to clopidogrel compared with patients randomized to prasugrel (rate ratio, 1.00; 95% CI, 0.99–1.01). There was no difference in DAOH between treatment arms in any subgroup, including the younger adult population, the older adult population, and the frail/prefrail population (Table 2). There was no association between treatment arm and DAOH in any geographic area (interaction P=0.39; Table III in the Data Supplement). When we counted DAOH from the time of randomization through 12 months (rather than from the time of discharge), there remained no difference in DAOH between treatment arms overall or in any subgroup (Table IV in the Data Supplement).

    Table 2. DAOH by Treatment Arm

    SubgroupMean DAOH (SD)Rate Ratio (95% CI)P Value
    Overall316 (87)317 (86)1.00 (0.99–1.01)0.67
    Younger adult (<75-y old)323 (79)324 (79)1.00 (0.99–1.01)0.57
    Older adult (≥75-y old)293 (106)293 (105)1.00 (0.98–1.03)0.87
    Frail301 (100)306 (100)1.01 (0.98–1.03)0.70

    DAOH indicates days alive and out of the hospital.


    In this analysis of TRILOGY ACS, undertaken to explore DAOH as an end point in a clinical trial comparing antithrombotic regimens in patients with ACS, we found that DAOH could be feasibly calculated using only site-submitted adverse event reporting data without relying on central adjudication. Older and frail patients had fewer DAOH than younger patients; patients with ischemic and bleeding events had fewer DAOH than those without. Nearly 30% of patients either died or were hospitalized over 12 months’ follow-up. There was no difference in DAOH between the prasugrel and clopidogrel arms in the overall population, younger adults, older adults, or frail patients, and across geographic regions.

    In its primary analysis, TRILOGY ACS found that treatment with prasugrel, as compared to clopidogrel, had no significant effect on either the primary efficacy end point of time to first incidence of cardiovascular death, MI, or stroke (hazard ratio, 0.91; 95% CI, 0.79–1.05) or the primary safety end point of TIMI major bleeding (hazard ratio, 1.31; 95% CI, 0.81–2.11).10 Our DAOH analysis thus complements the primary trial results, reframing them in a more patient-centered way.

    Cardiovascular clinical trials undertaken to evaluate new agents or new indications for existing medications uniformly have a primary end point of time to first incidence of a composite of ischemic events—most frequently, cardiovascular death, MI, and stroke.10,17–20 With increasing recognition of the prognostic importance of bleeding events,21 several recent trials have incorporated major bleeding as a primary safety end point or combined bleeding and ischemic events into a net clinical benefit or net adverse events end point.22,23 This approach has several limitations that make it less patient-centered and less translatable to pragmatic clinical trials.

    First, separate analysis of ischemic and bleeding events requires guideline-writing committees, practicing physicians, and regulatory agencies to make their own judgments on how to weight each type of event, and observational studies have shown that physician behavior is driven largely by safety considerations, rather than efficacy, potentially contributing to the risk-treatment paradox.24,25 Second, time-to-event analyses fail to account for the possibility of multiple events and thus do not use all of the available data generated during follow-up. In trials with composite end points, the time-to-event framework also means that patients suffering less severe events early during follow-up are judged to have worse outcomes than patients suffering more severe events later; for example, a patient with a minor, nonfatal MI 30 days into follow-up has a worse outcome under the time-to-event framework than a patient with a fatal MI at 31 days. Third, the clinical significance of composite ischemic events is uncertain, and neither ischemic events nor major bleeding themselves represents truly patient-centered outcomes: From the perspective of a patient, the importance of any diagnosis is not captured by the diagnosis alone but by the morbidity or mortality it causes.6 Lastly, the traditional time-to-composite-event framework requires a larger sample size compared with a continuous outcome and traditionally requires end point adjudication to determine whether nonfatal events occurred and cause of death for patients that died. End point adjudication and larger sample sizes increase trials’ cost and complexity, and as pragmatic, patient-centered trials become more common, these limitations may become more important.26

    A number of alternative frameworks have been proposed to circumvent some of the limitations of the traditional time-to-composite-ischemic-event framework; however, all of these frameworks have their own limitations. The net adverse events framework combines ischemic and bleeding events into a single composite end point.7 Though this approach avoids difficulties associated with integrating the results of separate efficacy and safety analyses, it does not take into account multiple events, requires end point adjudication, is not more patient-centered than the traditional time-to-event framework, and makes the a priori assumption that all ischemic and bleeding events are of equal importance, when this is clearly not the case.27 The recently completed PIONEER AF-PCI trial (Open-Label, Randomized, Controlled, Multicenter Study Exploring Two Treatment Strategies of Rivaroxaban and a Dose-Adjusted Oral Vitamin K Antagonist Treatment Strategy in Subjects With Atrial Fibrillation Who Undergo Percutaneous Coronary Intervention) included a secondary analysis with an end point of time to the composite of all-cause death or hospitalization.28 This strategy, which combines ischemic and bleeding events into a single, more patient-centered outcome, nevertheless does not take into account multiple events and assumes that readmission and death are events of equal severity.

    The Andersen-Gill method is one of several extensions to the traditional time-to-event framework that is able to account for the possibility that patients can have multiple events during follow-up.29,30 However, these time-to-recurrent-event approaches assume that all components of the composite end point are equally important, reducing their patient-centeredness. In fact, because patients can have multiple nonfatal events but only one fatal event, these approaches actually weight nonfatal events more heavily than fatal events.

    The weighted composite end point approach attempts to circumvent this limitation by prospectively assigning weights to various components of the composite end point that reflect the relative severity of each event.8,31 All patients start with a value of 1.0, and patients with nonfatal events are considered to have their contribution to the cohort reduced by some fraction depending on event severity. With multiple recurrent events, each patient’s contribution is further reduced, and with fatal events, the patient’s contribution goes to 0. A modified Kaplan-Meier curve can be plotted based on a life table generated from this approach. Though weights are assigned prospectively in this approach, and patient input on end point severity is often sought, assignment of weights to clinical events by a physician committee may not truly reflect the patient experience, and consensus-derived weights will not reflect the values and preferences of every patient to whom a clinician hopes to apply clinical trial results.

    The win ratio approach simplifies the abstract weighting of events by ranking them—for example, death is worse than stroke, which is worse than MI.32 Patients are then paired by baseline illness severity and treatment assignment (so that a pair contains 1 patient randomized to each treatment), and each pair is evaluated as to which member died first; the patient that died later (or not at all) wins that pair. If neither patient died, the process is repeated for each event type in order of declining severity. Ultimately, this approach yields a win ratio which can be thought of as the ratio of pairs in which the experimental treatment was the winner to pairs in which it was the loser. The win ratio approach is limited by its inability to account for event severity in a nuanced way, the difficulty in incorporating bleeding events into this framework, its need for end point adjudication, and the difficulty in understanding the clinical significance of its output.

    DAOH provides a simple alternative to the traditional time-to-composite-event framework, as well as the alternative frameworks. It accounts for multiple events and event severity, albeit in an indirect manner, with death generally contributing to a greater loss of DAOH than hospitalization. It estimates event severity with days hospitalized rather than assuming all events are of the same severity. It thus obviates questions related to the significance of composite events: Time spent in the hospital is an adverse event that patients wish to avoid; longer hospitalizations with greater morbidity are definitively worse than short hospitalizations; and long-term sequelae of clinical events ultimately manifest as recurrent hospitalization or death. The simplicity of DAOH also provides flexibility to account for bleeding and other severe treatment-related side effects that lead to hospitalization without the need for potentially controversial a priori assignment of weights. Lastly, DAOH incorporates cardiovascular events into a continuous measure, which could reduce sample size in clinical trials, and can be derived entirely from administrative data, making it a potentially attractive outcome for low-cost, pragmatic clinical trials.

    DAOH does have, however, a number of limitations. Though it is a reasonable measure for days of good health, it does not take into account nursing home or rehabilitation facility stays nor does it take into account how well the patient may feel while at home or the psychosocial and cultural mores that modulate patients’ experience of illness, which limits its ability to truly represent days of good health. This may be particularly relevant for older adults, in whom hospitalization frequently leads to short- or long-term functional decline, the severity of which is associated with but not determined by the severity of illness precipitating hospitalization.33–35 For example, a patient with a massive stroke leading to a 10-day hospitalization but no further sequela will have a better outcome under the DAOH framework than a patient with several bleeding hospitalizations that cumulatively result in >10 days hospitalized, although the first patient’s outcome is truly worse. In other cases, hospitalizations of the same length, such as for an MI and a bleeding event, could carry different long-term prognoses.36 Modifying clinical trial protocols to collect days spent at nonhospital inpatient facilities would only trivially increase trial cost and complexity, and a smartphone-based application could eventually enable patients to report daily symptom burden. Moreover, unlike the traditional-time-to-composite-ischemic-event framework, DAOH counts both cardiovascular and noncardiovascular events. Though this is a strength in the case of bleeding and other treatment-related side effects, it is generally not expected that cardiovascular medications will reduce noncardiovascular hospitalizations and deaths, and incorporation of these events into the DAOH framework may add statistical noise that obscures true treatment differences. In TRILOGY ACS, there were 3-fold as many hospitalizations as there were MIs, strokes, and major or minor bleeding events; however, this type of rough measure is likely to significantly undercount the number of hospitalizations owing to ischemic and bleeding events (and their sequela) because it does not account for heart failure hospitalizations, unstable angina, unplanned revascularizations, or bleeding-related hospitalizations that do not meet criteria for TIMI major or minor bleeding. Future studies should retrospectively examine treatment effects on DAOH in the context of clearly positive clinical trials to determine if DAOH’s incorporation of noncardiovascular events obscures treatment differences identified using a traditional time-to-composite-event framework. In addition, DAOH is limited by an inability to account for censoring at the end of the follow-up period; a patient who has an MI with a 3-day hospital stay during month 1 of follow-up and then remains healthy for the duration of follow-up will have fewer DAOH (a worse outcome) than a patient who remains healthy until dying at day 364. Importantly, however, the patient with early MI would have a worse outcome than the patient with late death under a traditional time-to-composite-event framework as well. Because the DAOH framework we used requires that most patients have complete follow-up through the end of the study period, it may be less useful in clinical trials evaluating agents used chronically. In such trials, the differential effect of the 2 treatment regimens may not be apparent until well after the median follow-up duration, and DAOH will not capture the treatment difference. Indeed, in TRILOGY ACS, there was a suggestion of a time-dependent benefit of prasugrel that emerged only after 12 months follow-up, which we were unable to explore in this analysis. In addition, the distribution of DAOH is highly skewed and influenced by the vast majority of patients who have no events during follow-up. This limitation could potentially be avoided by including DAOH as part of a stepwise analytic framework, in which investigators sequentially analyze the effect of a treatment on overall survival, then time to first hospitalization, and finally DAOH among patients that were hospitalized or died. Lastly, though DAOH seems to be an easily understandable and patient-centered outcome, future studies should examine patients’ ability to understand different presentations of clinical trial results, including both traditional time-to-composite-event presentations and alternate presentations, like DAOH. No literature yet exists to determine the clinical relevance of a given difference in DAOH between treatments; however, existing cost-effectiveness literature could be adapted for this purpose, taking into account both the financial costs of hospitalization, including resource intensity of hospital days, and quality of life.


    In TRILOGY ACS, DAOH could be feasibly calculated using only site-submitted adverse event data. Treatment with prasugrel, as compared with clopidogrel, had no effect on DAOH over 12-month follow-up, paralleling the main trial results. These findings support the utility of DAOH as a more patient-centered clinical trial outcome, which incorporates multiple events over the course of follow-up, balances putative reductions in ischemic events with increases in bleeding events, and can be determined without a formal adjudication process. Given these advantages, it may be reasonable for investigators to consider DAOH for inclusion in future clinical trials as a primary or key secondary outcome.


    The Data Supplement is available at

    Alexander Fanaroff, MD, Duke Clinical Research Institute, Duke University, 2400 Pratt St, Durham, NC 27705. Email


    • 1. Fanaroff AC, Hasselblad V, Roe MT, Bhatt DL, James SK, Steg PG, Gibson CM, Ohman EM. Antithrombotic agents for secondary prevention after acute coronary syndromes: a systematic review and network meta-analysis.Int J Cardiol. 2017; 241:87–96. doi: 10.1016/j.ijcard.2017.03.046CrossrefMedlineGoogle Scholar
    • 2. Bavishi C, Panwar S, Messerli FH, Bangalore S. Meta-analysis of comparison of the newer oral P2Y12 inhibitors (prasugrel or ticagrelor) to clopidogrel in patients with non-ST-elevation acute coronary syndrome.Am J Cardiol. 2015; 116:809–817. doi: 10.1016/j.amjcard.2015.05.058CrossrefMedlineGoogle Scholar
    • 3. Oldgren J, Wallentin L, Alexander JH, James S, Jönelid B, Steg G, Sundström J. New oral anticoagulants in addition to single or dual antiplatelet therapy after an acute coronary syndrome: a systematic review and meta-analysis.Eur Heart J. 2013; 34:1670–1680. doi: 10.1093/eurheartj/eht049CrossrefMedlineGoogle Scholar
    • 4. Amsterdam EA, Wenger NK, Brindis RG, Casey DE, Ganiats TG, Holmes DR, Jaffe AS, Jneid H, Kelly RF, Kontos MC, Levine GN, Liebson PR, Mukherjee D, Peterson ED, Sabatine MS, Smalling RS, Zieman SJ. 2014 AHA/ACC guideline for the management of patients with non-ST-elevation acute coronary syndromes.Circulation. 2014; 130:e344–e426. doi: 10.1161/CIR.0000000000000134LinkGoogle Scholar
    • 5. Roffi M, Patrono C, Collet JP, Mueller C, Valgimigli M, Andreotti F, Bax JJ, Borger MA, Brotons C, Chew DP, Gencer B, Hasenfuss G, Kjeldsen K, Lancellotti P, Landmesser U, Mehilli J, Mukherjee D, Storey RF, Windecker S; ESC Scientific Document Group. 2015 ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation: task force for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation of the European Society of Cardiology (ESC).Eur Heart J. 2016; 37:267–315. doi: 10.1093/eurheartj/ehv320CrossrefMedlineGoogle Scholar
    • 6. Verheugt FW, Clemmensen P, Mehran R, Agewall S, Pocock SJ, Goldstein S, Torp-Pedersen C, Simoons ML, Borer JS, Khder YM, Burton P, Deliargyris E, McMurray JJ, Berkowitz SD, Stough WG, Zannad F. Antithrombotic outcome trials in acute coronary syndromes: seeking the optimal balance between safety and efficacy.Eur Heart J. 2013; 34:1621–1629. doi: 10.1093/eurheartj/eht013CrossrefMedlineGoogle Scholar
    • 7. Valgimigli M, Frigoli E, Leonardi S, Rothenbühler M, Gagnor A, Calabrò P, Garducci S, Rubartelli P, Briguori C, Andò G, Repetto A, Limbruno U, Garbo R, Sganzerla P, Russo F, Lupi A, Cortese B, Ausiello A, Ierna S, Esposito G, Presbitero P, Santarelli A, Sardella G, Varbella F, Tresoldi S, de Cesare N, Rigattieri S, Zingarelli A, Tosi P, van ‘t Hof A, Boccuzzi G, Omerovic E, Sabaté M, Heg D, Jüni P, Vranckx P; MATRIX Investigators. Bivalirudin or unfractionated heparin in acute coronary syndromes.N Engl J Med. 2015; 373:997–1009. doi: 10.1056/NEJMoa1507854CrossrefMedlineGoogle Scholar
    • 8. Bakal JA, Roe MT, Ohman EM, Goodman SG, Fox KA, Zheng Y, Westerhout CM, Hochman JS, Lokhnygina Y, Brown EB, Armstrong PW. Applying novel methods to assess clinical outcomes: insights from the TRILOGY ACS trial.Eur Heart J. 2015; 36:385–92a. doi: 10.1093/eurheartj/ehu262CrossrefMedlineGoogle Scholar
    • 9. Ariti CA, Cleland JG, Pocock SJ, Pfeffer MA, Swedberg K, Granger CB, McMurray JJ, Michelson EL, Ostergren J, Yusuf S. Days alive and out of hospital and the patient journey in patients with heart failure: Insights from the candesartan in heart failure: assessment of reduction in mortality and morbidity (CHARM) program.Am Heart J. 2011; 162:900–906. doi: 10.1016/j.ahj.2011.08.003CrossrefMedlineGoogle Scholar
    • 10. Roe MT, Armstrong PW, Fox KA, White HD, Prabhakaran D, Goodman SG, Cornel JH, Bhatt DL, Clemmensen P, Martinez F, Ardissino D, Nicolau JC, Boden WE, Gurbel PA, Ruzyllo W, Dalby AJ, McGuire DK, Leiva-Pons JL, Parkhomenko A, Gottlieb S, Topacio GO, Hamm C, Pavlides G, Goudev AR, Oto A, Tseng CD, Merkely B, Gasparovic V, Corbalan R, Cinteză M, McLendon RC, Winters KJ, Brown EB, Lokhnygina Y, Aylward PE, Huber K, Hochman JS, Ohman EM; TRILOGY ACS Investigators. Prasugrel versus clopidogrel for acute coronary syndromes without revascularization.N Engl J Med. 2012; 367:1297–1309. doi: 10.1056/NEJMoa1205512CrossrefMedlineGoogle Scholar
    • 11. Chin CT, Roe MT, Fox KA, Prabhakaran D, Marshall DA, Petitjean H, Lokhnygina Y, Brown E, Armstrong PW, White HD, Ohman EM; TRILOGY ACS Steering Committee. Study design and rationale of a comparison of prasugrel and clopidogrel in medically managed patients with unstable angina/non-ST-segment elevation myocardial infarction: the TaRgeted platelet Inhibition to cLarify the Optimal strateGy to medicallY manage Acute Coronary Syndromes (TRILOGY ACS) trial.Am Heart J. 2010; 160:16–22.e1. doi: 10.1016/j.ahj.2010.04.022CrossrefMedlineGoogle Scholar
    • 12. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, Seeman T, Tracy R, Kop WJ, Burke G, McBurnie MA; Cardiovascular Health Study Collaborative Research Group. Frailty in older adults: evidence for a phenotype.J Gerontol A Biol Sci Med Sci. 2001; 56:M146–M156.CrossrefMedlineGoogle Scholar
    • 13. White HD, Westerhout CM, Alexander KP, Roe MT, Winters KJ, Cyr DD, Fox KA, Prabhakaran D, Hochman JS, Armstrong PW, Ohman EM; TRILOGY ACS Investigators. Frailty is associated with worse outcomes in non-ST-segment elevation acute coronary syndromes: insights from the TaRgeted platelet Inhibition to cLarify the Optimal strateGy to medicallY manage Acute Coronary Syndromes (TRILOGY ACS) trial.Eur Heart J Acute Cardiovasc Care. 2016; 5:231–242. doi: 10.1177/2048872615581502CrossrefMedlineGoogle Scholar
    • 14. Brännäs K. Limited dependent poisson regression.The Statistician. 1992; 41:413–423.CrossrefGoogle Scholar
    • 15. Levine GN, Bates ER, Bittl JA, Brindis RG, Fihn SD, Fleisher LA, Granger CB, Lange RA, Mack MJ, Mauri L, Mehran R, Mukherjee D, Newby LK, O’Gara PT, Sabatine MS, Smith PK, Smith SC. 2016 ACC/AHA guideline focused update on duration of dual antiplatelet therapy in patients with coronary artery disease: a report of the American College of Cardiology/American Heart Association task force on clinical practice guidelines: an update of the 2011 ACCF/AHA/SCAI guideline for percutaneous coronary intervention, 2011 ACCF/AHA guideline for coronary artery bypass graft surgery, 2012 ACC/AHA/ACP/AATS/PCNA/SCAI/STS guideline for the diagnosis and management of patients with stable ischemic heart disease, 2013 ACCF/AHA guideline for the management of ST-elevation myocardial infarction, 2014 AHA/ACC guideline for the management of patients with non-ST-elevation acute coronary syndromes, and 2014 ACC/AHA guideline on perioperative cardiovascular evaluation and management of patients undergoing noncardiac surgery.Circulation. 2016; 134:e123–e155. doi: 10.1161/CIR.0000000000000404LinkGoogle Scholar
    • 16. O’Gara PT, Kushner FG, Ascheim DD, Casey DE, Chung MK, de Lemos JA, Ettinger SM, Fang JC, Fesmire FM, Franklin BA, Granger CB, Krumholz HM, Linderbaum JA, Morrow DA, Newby LK, Ornato JP, Ou N, Radford MJ, Tamis-Holland JE, Tommaso CL, Tracy CM, Woo YJ, Zhao DX, Anderson JL, Jacobs AK, Halperin JL, Albert NM, Brindis RG, Creager MA, DeMets D, Guyton RA, Hochman JS, Kovacs RJ, Kushner FG, Ohman EM, Stevenson WG, Yancy CW; American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. 2013 ACCF/AHA guideline for the management of ST-elevation myocardial infarction: a report of the American College of Cardiology Foundation/American Heart Association task force on practice guidelines.Circulation. 2013; 127:e362–e425. doi: 10.1161/CIR.0b013e3182742cf6LinkGoogle Scholar
    • 17. Wiviott SD, Braunwald E, McCabe CH, Montalescot G, Ruzyllo W, Gottlieb S, Neumann FJ, Ardissino D, De Servi S, Murphy SA, Riesmeyer J, Weerakkody G, Gibson CM, Antman EM; TRITON-TIMI 38 Investigators. Prasugrel versus clopidogrel in patients with acute coronary syndromes.N Engl J Med. 2007; 357:2001–2015. doi: 10.1056/NEJMoa0706482CrossrefMedlineGoogle Scholar
    • 18. Wallentin L, Becker RC, Budaj A, Cannon CP, Emanuelsson H, Held C, Horrow J, Husted S, James S, Katus H, Mahaffey KW, Scirica BM, Skene A, Steg PG, Storey RF, Harrington RA, Freij A, Thorsén M; PLATO Investigators. Ticagrelor versus clopidogrel in patients with acute coronary syndromes.N Engl J Med. 2009; 361:1045–1057. doi: 10.1056/NEJMoa0904327CrossrefMedlineGoogle Scholar
    • 19. Mega JL, Braunwald E, Wiviott SD, Bassand JP, Bhatt DL, Bode C, Burton P, Cohen M, Cook-Bruns N, Fox KA, Goto S, Murphy SA, Plotnikov AN, Schneider D, Sun X, Verheugt FW, Gibson CM; ATLAS ACS 2–TIMI 51 Investigators. Rivaroxaban in patients with a recent acute coronary syndrome.N Engl J Med. 2012; 366:9–19. doi: 10.1056/NEJMoa1112277CrossrefMedlineGoogle Scholar
    • 20. Tricoci P, Huang Z, Held C, Moliterno DJ, Armstrong PW, Van de Werf F, White HD, Aylward PE, Wallentin L, Chen E, Lokhnygina Y, Pei J, Leonardi S, Rorick TL, Kilian AM, Jennings LH, Ambrosio G, Bode C, Cequier A, Cornel JH, Diaz R, Erkan A, Huber K, Hudson MP, Jiang L, Jukema JW, Lewis BS, Lincoff AM, Montalescot G, Nicolau JC, Ogawa H, Pfisterer M, Prieto JC, Ruzyllo W, Sinnaeve PR, Storey RF, Valgimigli M, Whellan DJ, Widimsky P, Strony J, Harrington RA, Mahaffey KW; TRACER Investigators. Thrombin-receptor antagonist vorapaxar in acute coronary syndromes.N Engl J Med. 2012; 366:20–33. doi: 10.1056/NEJMoa1109719CrossrefMedlineGoogle Scholar
    • 21. Rao SV, O’Grady K, Pieper KS, Granger CB, Newby LK, Van de Werf F, Mahaffey KW, Califf RM, Harrington RA. Impact of bleeding severity on clinical outcomes among patients with acute coronary syndromes.Am J Cardiol. 2005; 96:1200–1206. doi: 10.1016/j.amjcard.2005.06.056CrossrefMedlineGoogle Scholar
    • 22. Bonaca MP, Bhatt DL, Cohen M, Steg PG, Storey RF, Jensen EC, Magnani G, Bansilal S, Fish MP, Im K, Bengtsson O, Oude Ophuis T, Budaj A, Theroux P, Ruda M, Hamm C, Goto S, Spinar J, Nicolau JC, Kiss RG, Murphy SA, Wiviott SD, Held P, Braunwald E, Sabatine MS; PEGASUS-TIMI 54 Steering Committee and Investigators. Long-term use of ticagrelor in patients with prior myocardial infarction.N Engl J Med. 2015; 372:1791–1800. doi: 10.1056/NEJMoa1500857CrossrefMedlineGoogle Scholar
    • 23. Mauri L, Kereiakes DJ, Yeh RW, Driscoll-Shempp P, Cutlip DE, Steg PG, Normand SL, Braunwald E, Wiviott SD, Cohen DJ, Holmes DR, Krucoff MW, Hermiller J, Dauerman HL, Simon DI, Kandzari DE, Garratt KN, Lee DP, Pow TK, Ver Lee P, Rinaldi MJ, Massaro JM; DAPT Study Investigators. Twelve or 30 months of dual antiplatelet therapy after drug-eluting stents.N Engl J Med. 2014; 371:2155–2166. doi: 10.1056/NEJMoa1409312CrossrefMedlineGoogle Scholar
    • 24. Beigel R, Iakobishvili Z, Shlomo N, Segev A, Witberg G, Zahger D, Atar S, Alcalai R, Kapeliovich M, Gottlieb S, Goldenberg I, Asher E, Matetzky S. Real-world use of novel P2Y12 inhibitors in patients with acute myocardial infarction: a treatment paradox.Cardiology. 2017; 136:21–28. doi: 10.1159/000447396CrossrefMedlineGoogle Scholar
    • 25. Beigel R, Iakobishvili Z, Shlomo N, Segev A, Witberg G, Zahger D, Atar S, Alcalai R, Kapeliovich M, Gottlieb S, Goldenberg I, Asher E, Matetzky S. Real-world use of novel P2Y12 inhibitors in patients with acute myocardial infarction: a treatment paradox.Cardiology. 2017; 136:21–28. doi: 10.1159/000447396CrossrefMedlineGoogle Scholar
    • 26. Jones WS, Roe MT, Antman EM, Pletcher MJ, Harrington RA, Rothman RL, Oetgen WJ, Rao SV, Krucoff MW, Curtis LH, Hernandez AF, Masoudi FA. The changing landscape of randomized clinical trials in cardiovascular disease.J Am Coll Cardiol. 2016; 68:1898–1907. doi: 10.1016/j.jacc.2016.07.781CrossrefMedlineGoogle Scholar
    • 27. Steg PG, Bhatt DL. Is there really a benefit to net clinical benefit in testing antithrombotics?Circulation. 2018; 137:1429–1431. doi: 10.1161/CIRCULATIONAHA.117.033442LinkGoogle Scholar
    • 28. Gibson CM, Pinto DS, Chi G, Arbetter D, Yee M, Mehran R, Bode C, Halperin J, Verheugt FW, Wildgoose P, Burton P, van Eickels M, Korjian S, Daaboul Y, Jain P, Lip GY, Cohen M, Peterson ED, Fox KA. Recurrent hospitalization among patients with atrial fibrillation undergoing intracoronary stenting treated with 2 treatment strategies of rivaroxaban or a dose-adjusted oral vitamin K antagonist treatment strategy.Circulation. 2017; 135:323–333. doi: 10.1161/CIRCULATIONAHA.116.025783LinkGoogle Scholar
    • 29. Andersen PK, Gill RD. Cox’s regression model for counting processes: a large sample study.Ann Stat. 1982; 10:1100–1120.CrossrefGoogle Scholar
    • 30. Amorim LD, Cai J. Modelling recurrent events: a tutorial for analysis in epidemiology.Int J Epidemiol. 2015; 44:324–333. doi: 10.1093/ije/dyu222CrossrefMedlineGoogle Scholar
    • 31. Bakal JA, Westerhout CM, Cantor WJ, Fernández-Avilés F, Welsh RC, Fitchett D, Goodman SG, Armstrong PW. Evaluation of early percutaneous coronary intervention vs. standard therapy after fibrinolysis for ST-segment elevation myocardial infarction: contribution of weighting the composite endpoint.Eur Heart J. 2013; 34:903–908. doi: 10.1093/eurheartj/ehs438CrossrefMedlineGoogle Scholar
    • 32. Pocock SJ, Ariti CA, Collier TJ, Wang D. The win ratio: a new approach to the analysis of composite end points in clinical trials based on clinical priorities.Eur Heart J. 2012; 33:176–182. doi: 10.1093/eurheartj/ehr352CrossrefMedlineGoogle Scholar
    • 33. Mehta KM, Pierluissi E, Boscardin WJ, Kirby KA, Walter LC, Chren MM, Palmer RM, Counsell SR, Landefeld CS. A clinical index to stratify hospitalized older adults according to risk for new-onset disability.J Am Geriatr Soc. 2011; 59:1206–1216. doi: 10.1111/j.1532-5415.2011.03409.xCrossrefMedlineGoogle Scholar
    • 34. Fimognari FL, Pierantozzi A, De Alfieri W, Salani B, Zuccaro SM, Arone A, Palleschi G, Palleschi L. The severity of acute illness and functional trajectories in hospitalized older medical patients.J Gerontol A Biol Sci Med Sci. 2017; 72:102–108.CrossrefMedlineGoogle Scholar
    • 35. Volpato S, Onder G, Cavalieri M, Guerra G, Sioulis F, Maraldi C, Zuliani G, Fellin R; Italian Group of Pharmacoepidemiology in the Elderly Study (GIFA). Characteristics of nondisabled older patients developing new disability associated with medical illnesses and hospitalization.J Gen Intern Med. 2007; 22:668–674. doi: 10.1007/s11606-007-0152-1CrossrefMedlineGoogle Scholar
    • 36. Valgimigli M, Costa F, Lokhnygina Y, Clare RM, Wallentin L, Moliterno DJ, Armstrong PW, White HD, Held C, Aylward PE, Van de Werf F, Harrington RA, Mahaffey KW, Tricoci P. Trade-off of myocardial infarction vs. bleeding types on mortality after acute coronary syndrome: lessons from the Thrombin Receptor Antagonist for Clinical Event Reduction in Acute Coronary Syndrome (TRACER) randomized trial.Eur Heart J. 2017; 38:804–810. doi: 10.1093/eurheartj/ehw525MedlineGoogle Scholar