Rates of clinical outcomes through 90 days according to age. Differences in rates of each outcome across age strata were significant at P < .001. CHF indicates congestive heart failure.
Kaplan-Meier estimates of 90-day mortality according to age. Differences in mortality rates by age were significant at P < .001.
Estimates of adjusted mortality rates. A, Rates within the first week after primary percutaneous coronary intervention according to age (P < .001 for comparisons by age). B, Estimates within 90 days of primary percutaneous coronary intervention according to age (P < .001 for comparisons by age).
Gharacholou SM, Lopes RD, Alexander KP, Mehta RH, Stebbins AL, Pieper KS, James SK, Armstrong PW, Granger CB. Age and Outcomes in ST-Segment Elevation Myocardial Infarction Treated With Primary Percutaneous Coronary InterventionFindings From the APEX-AMI Trial. Arch Intern Med. 2011;171(6):559-567. doi:10.1001/archinternmed.2011.36
Copyright 2011 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2011
To understand the influence of age on treatment and outcomes, we analyzed the largest group of patients 75 years or older with ST-segment elevation myocardial infarction treated with primary percutaneous coronary intervention (PPCI) in a clinical trial.
We analyzed data from 5745 patients in the Assessment of Pexelizumab in Acute Myocardial Infarction trial from July 13, 2004, through May 11, 2006. Age was analyzed continuously and according to 3 groups: younger than 65 years (n = 3410), 65 to 74 years old (n = 1358), and 75 years or older (n = 977). The main outcome measures were 90-day mortality and the composite of congestive heart failure, shock, or death at 90 days.
Older patients had higher rates of hypertension, chronic obstructive lung disease, previous angina, and prior revascularization. Also notable in these patients were higher Killip class, less angiographic success after PPCI, and less ST-segment resolution with higher rates of in-hospital clinical events, including mechanical, electrical, and bleeding complications. There was less use of short-term adjunctive medications but similar use of discharge medications in older compared with younger patients. Ninety-day mortality rates were 2.3%, 4.8%, and 13.1%; composite outcome rates were 5.9%, 11.9%, and 22.8% for patients younger than 65 years, 65 to 74 years old, and 75 years or older, respectively. After multivariable adjustment, age was the strongest independent predictor of 90-day mortality (hazard ratio, 2.07 per 10-year increase; 95% confidence interval, 1.84-2.33).
Older patients have lower rates of acute procedural success and more postinfarction complications. Age is the strongest predictor of 90-day mortality in ST-segment elevation myocardial infarction patients undergoing PPCI. Despite implementing PPCI for ST-segment elevation myocardial infarction in older patients, early risk remains high, necessitating continued focus on improving outcomes in this vulnerable population.
Among patients with ST-segment elevation myocardial infarction (STEMI), observational registries have shown that 30% are 75 years or older.1,2 Primary percutaneous coronary intervention (PPCI) is the preferred reperfusion strategy for older patients with STEMI given their greater baseline ischemic and hemorrhagic risk, presentation delay, and greater relative contraindications to pharmacologic-based reperfusion.2- 10 However, prior observations from clinical trials of PPCI in older patients (≥75 years) with STEMI have been based on relatively small numbers of participants.11- 13 Low representation of older STEMI patients in PPCI trials has limited a full evaluation of outcomes considering guideline-recommended use of pharmacologic therapies,2,3 such as β-blockers,14,15 angiotensin-converting enzyme inhibitors,16 and statins.17- 19 In addition, procedural characteristics by age in STEMI patients treated with PPCI have not been well described.20- 22 Higher mortality rates in older STEMI patients22,23 may be related to a higher incidence of post–myocardial infarction (MI) complications, such as congestive heart failure (CHF), stroke, bleeding, and renal failure.1,21 Finally, the multidimensional health status of older patients with STEMI adds additional complexity.2
To better understand the relationship between age and outcomes, we analyzed 5745 patients with STEMI undergoing PPCI, including 977 patients 75 years or older, from the Assessment of Pexelizumab in Acute Myocardial Infarction (APEX-AMI) trial.24,25 Our principal objectives were to assess the association between age and clinical and procedural characteristics, in-hospital complications, and cardiovascular events through 90 days and to evaluate the independent contribution of age in predicting 90-day mortality in this population.
The details of the rationale, design, and findings of the APEX-AMI trial have been previously published.24,25 Briefly, APEX-AMI was a multicenter, randomized, double-blind, placebo-controlled trial comparing intravenous (IV) pexelizumab (an inhibitor of complement) with placebo in 5745 electrocardiographically high-risk STEMI patients expected to undergo PPCI (93.6% of patients underwent PPCI, and the analysis was based on intent to treat) as their clinically intended reperfusion strategy. A high-risk electrocardiogram (ECG) was defined as an anterior MI, presumed new left bundle branch block with 1-mm concordant ST-segment elevation, or high-risk inferior MI (8 mm of total ST-segment deviation). Patients were eligible for the study if they were 18 years or older, had experienced symptoms for at least 20 minutes, and had presented with STEMI within 6 hours of symptom onset. Patients underwent randomization before angiography. Exclusions to study participation included prior treatment with fibrinolytic therapy, isolated inferior MI, pregnancy or breastfeeding, complement deficiency, or active serious infection. Otherwise, high-risk patients were encouraged to be enrolled, including patients with cardiogenic shock and renal failure. There was no upper age exclusion to trial participation. The trial was conducted from July 13, 2004, through May 11, 2006, in 17 countries and included 296 sites. The primary study end point was all-cause 30-day mortality, and secondary study end points included the composite incidence of CHF, cardiogenic shock, or death through days 30 and 90.
The clinical outcome of CHF included new or worsening CHF or subsequent hospitalization for CHF. New CHF was defined as occurring during the index hospitalization and was required to begin or persist for more than 24 hours after randomization. Worsening CHF was defined as occurring or persisting more than 24 hours after randomization. Congestive heart failure was defined on the basis of the physician's clinical decision to treat the patient with an IV diuretic, IV inotropic agent, or IV vasodilator and at least 1 of the following: presence of pulmonary edema on chest radiograph believed to be of cardiac origin, rales more than one-third up the lung fields and believed to be due to CHF, pulmonary capillary wedge pressure or left ventricular end diastolic pressure greater than 18 mm Hg, or dyspnea with documented PO2 less than 80 mm Hg on room air or oxygen saturation less than 90% on room air without significant lung disease. Subsequent hospitalization for CHF was defined as readmission to an acute care facility primarily for the treatment of CHF and was required to include IV treatment with a diuretic, inotrope, or vasodilator. Cardiogenic shock was defined as systolic blood pressure of 90 mm Hg lasting for at least 1 hour, not responsive to fluid resuscitation and/or heart rate correction, believed to be secondary to cardiac dysfunction, and associated with at least 1 of the following signs of hypoperfusion: cool, clammy skin; oliguria; altered sensorium; and/or cardiac index of 2.2 L/min/m2 or less. Primary and secondary end points were adjudicated by a clinical events committee masked to treatment arm based on previously agreed-on end point definitions.24- 26 Recurrent MI within 24 hours of randomization was defined as reelevation of creatine kinase MB by 33% or more from the preceding value or 100% from the preceding nadir and that reached a level of at least 3 times the upper limit of normal with accompanying ischemic symptoms. Recurrent MI 24 hours after randomization was defined as new pathologic Q waves, reelevation of creatine kinase MB to at least 3 times (at 24 hours to discharge) or at least 2 times the upper limit of normal (after hospital discharge), and accompanying ischemic symptoms. ST-segment resolution was assessed by the ECG core laboratories by comparing the ECG 30 minutes after PCI with the baseline ECG and defined as the following: complete ST-segment resolution (≥70%), partial resolution (30%-70%), and no resolution (<30%), as has been previously reported.27 Procedural success was defined as final Thrombolysis in Myocardial Infarction (TIMI) flow grade 2 or 3 (site-reported) in the infarct-related artery. Data were collected using an electronic case report form. In-hospital medication use and medication use at discharge were also obtained. In the APEX-AMI trial, there were no meaningful differences in any of the major outcomes between the pexelizumab group and the placebo group; therefore, both treatment arms were pooled for the current analysis.
We included all patients from the APEX-AMI study in the current analysis to understand the association of age with treatment patterns, in-hospital events, and clinical outcomes. All institutional review boards of participating medical centers approved the APEX-AMI protocol, and all study participants provided written informed consent. The present analysis was approved by the institutional review board of Duke University Medical Center.
Age was analyzed as a continuous variable for modeling purposes and reported categorically for descriptive purposes. Baseline clinical characteristics were reported by the following age groups: younger than 65 years, 65 to 74 years old, and 75 years or older. Statistical comparisons are reported on patients younger than 75 years and 75 years or older. Continuous variables are expressed as medians with 25th and 75th percentiles and categorical variables as counts and percentages. Univariate analyses for continuous variables were performed using Wilcoxon rank sum tests, whereas categorical variables were compared using χ2 or Fisher exact tests. Statistical significance was determined at the 2-sided α = .05 level. Kaplan-Meier curves and rates were generated to report adjusted and unadjusted 90-day outcomes by age.
Multivariable Cox proportional hazards models were used to develop a predictive model for 90-day mortality. Prespecified candidate variables used in model building included the following baseline characteristics at the time of randomization: age, sex, height, weight, body mass index, race, baseline serum creatinine level, systolic blood pressure, diastolic blood pressure, heart rate, Killip class, total ST-segment change, smoking status, MI location (high-risk inferior vs other), transfer status, randomized treatment (pexelizumab vs placebo), and country of enrollment. Candidate variables also included medical history, such as hypertension, diabetes mellitus, hyperlipidemia, MI, coronary artery disease, CHF, angina, PCI, coronary artery bypass graft surgery, chronic obstructive pulmonary disease, stroke, family history of coronary artery disease, current renal dialysis, chronic liver disease or inflammatory condition, history of peripheral vascular disease, and atrial fibrillation. Given that patients in the analysis were expected to undergo PPCI as the method of reperfusion, we included time from symptom onset to hospital admission, time from hospital admission to randomization, and time from symptom onset to randomization. Modeling assumptions were reviewed and transformations were performed as needed. Heart rate was modeled linearly with the point of inflection occurring at 70 per minute or less and greater than 70 per minute. Time from symptom onset to hospital arrival and time from hospital arrival to randomization were modeled as a linear spline. Backward, forward, and stepwise methods were implemented to attain variables statistically significant at the .05 level. Bootstrapping techniques were performed to assess model stability and for internal validation of model results. Using the total analysis population, 200 samples were drawn at random, with replacement, to determine whether variables were stable in predicting the outcome of 90-day mortality; only variables that remained stable across samples at least 75% of the time were accepted. All analyses were performed using SAS statistical software (version 8.2; SAS Institute Inc, Cary, North Carolina).
Of the 5745 patients enrolled in the APEX-AMI trial, 3410 were younger than 65 years, 1358 were 65 to 74 years, and 977 were 75 years or older. Of those 75 years or older, 455 patients were 80 years or older and 126 patients were 85 years or older. The oldest patient in the APEX-AMI trial was 97 years old. Patients 75 years or older accounted for 17.0% of the trial population. Older patients were enrolled in similar proportions within and outside the United States. The proportion of women increased across age groups, comprising 16.2% of patients younger than 65 years, 26.5% of patients 65 to 74 years old, and 42.4% of patients 75 years or older. Patients 75 years or older were more likely to have a history of hypertension, angina, prior revascularization (coronary artery bypass graft), CHF, and chronic obstructive pulmonary disease; they were also more likely to present with higher Killip class compared with patients younger than 75 years (Table 1). The median time from symptoms to randomization increased with age, from 2.7 hours in patients younger than 65 years to 3.1 hours in patients 75 years or older. Left bundle branch block was more common in older patients. There was no observed difference in location of MI by age strata.
Overall rates of in-hospital medication treatments, including aspirin, thienopyridines, glycoprotein IIb/IIIa inhibitors, β-blockers, and statins were slightly lower in patients 75 years or older compared with patients younger than 75 years (Table 2). Similar trends in overall medication use at hospital discharge in survivors of STEMI were observed across age groups and among those without contraindications; older patients were not less likely to receive discharge medications. Rates of thienopy-ridine use at discharge not excluding patients with contra indications, were high and similar in patients younger than 65 years (95.2%), 65 to 74 years (92.7%), and 75 years or older (89.6%). There were high rates of intracoronary stenting in trial participants (89.2%). Compared with those younger than 75 years, patients 75 years or older had similar median times until hospital arrival for PPCI (1.7 vs 1.6 hours) but lower rates of postprocedural TIMI grade 3 flow and less ST-segment resolution after PPCI despite similar infarct size (Table 3).
In-hospital clinical events increased with age (Table 4). The rates of electrical and mechanical complications, renal failure, hypotension, stroke, recurrent ischemic events, and bleeding occurred more often in PPCI-treated patients 75 years or older compared with PPCI-treated patients younger than 75 years. Rates of ventricular tachycardia or fibrillation, pericarditis, and pulmonary embolism were not significantly different in PPCI-treated patients 75 years or older compared with PPCI-treated patients younger than 75 years. Differences in clinical events by age in non–PPCI-treated patients were not as apparent and likely related to the smaller numbers of patients in this group (n = 369); however, clinical events occurred more often in non–PPCI-treated patients compared with PPCI-treated patients within each age strata. Stroke rates at 90 days in patients younger than 65 years, 65 to 74 years old, and 75 years or older were 0.7%, 1.5%, and 2.8%, respectively. There were 72 patients (1.3%) who had stroke within 90 days after randomization. Of the 69 patients who underwent PPCI and developed stroke, 18 (26.1%) developed stroke within 24 hours of PPCI. The median number of days from PPCI to stroke was 4.1,28 In the total APEX-AMI cohort, stroke occurring within 7 days of randomization occurred in 0.4%, 1.1%, and 1.5% of patients younger than 65 years, 65 to 74 years, and 75 years or older, respectively (P < .001).
Rates of CHF in patients younger than 65 years, 65 to 74 years, and 75 years or older were 2.7%, 5.8%, and 11.5%, respectively, and rates of shock were 2.1%, 4.3%, and 6.9%, respectively (Figure 1). Rates of the composite outcome (CHF, shock, or death) occurred in 5.9%, 11.9%, and 22.8% of patients younger than 65 years, 65 to 74 years, and 75 years or older, respectively. Rates of recurrent in-hospital MI were 2.6%, 4.1%, and 4.8% in patients younger than 65 years, 65 to 74 years, and 75 years or older, respectively. In the total APEX-AMI cohort, thirty-day mortality rates in patients younger than 65 years, 65 to 74 years, and 75 years or older were 2.0%, 4.1%, 10.8%, respectively; 90-day mortality rates across the same age strata were 2.3%, 4.8%, and 13.1%, respectively. When the analysis was restricted to non–PPCI-treated patients had 30-day mortality rates of 4.6%, 6.3%, and 20.8% in those younger than 65 years, 65 to 74 years, and 75 years or older, respectively. Despite an average 0.5% change in the mortality rate from 30 to 90 days in patients younger than 75 years, the mortality rate for patients 75 years or older steadily increased (2.3% change) during the same period in the APEX-AMI cohort (Figure 2). These data suggest that older STEMI patients treated with PPCI compared with patients younger than 75 years incur an early risk after PPCI and remain at higher risk for mortality and subsequent events in the early months after PPCI for STEMI.
The predictors of 90-day mortality using baseline characteristics were evaluated in multivariable analysis (Table 5). After adjustment, age was the strongest independent predictor of mortality, such that for every decade increase in age there was a 2-fold increase in the hazard of 90-day mortality (hazard ratio, 2.07 per 10-year increase; 95% confidence interval, 1.84-2.33). When explored by age groups, the adjusted hazard ratio for 90-day mortality was more than 5-fold higher in patients 75 years or older compared with patients younger than 65 years. Killip class of 3 or higher was also independently associated with a more than 4-fold increased hazard of 90-day mortality. Other factors associated with mortality were heart rate, serum creatinine level, ST-segment deviation, and high-risk inferior MI. After adjustment for these baseline factors, Kaplan-Meier mortality estimates illustrate the strong relationship between age and mortality in the APEX-AMI cohort in the early period after PPCI (Figure 3A) and months after treatment with PPCI for STEMI (Figure 3B).
In this study, we present the characteristics and outcomes of the largest number of STEMI patients 75 years or older (n = 977) who underwent PPCI in a randomized clinical trial. We observed that older patients compared with their younger counterparts achieved lower rates of TIMI grade 3 flow and ST-segment resolution postprocedurally, the latter having been previously shown to have independent prognostic significance even among patients with angiographically successful results.27 Older patients compared with younger patients experienced more complications after PPCI for STEMI. Despite high use of evidence-based medications and previous observations that optimal clinical practices occur in clinical trials,29 we observed that age remains the most powerful determinant of outcome, such that for every decade increase in age, there was a more than 2-fold increase in the adjusted hazard of 90-day mortality. Patients 75 years or older had 30- and 90-day mortality rates of 10.8% and 13.1%, respectively, including high rates of the 90-day composite outcome (CHF, shock, or death [22.8%]). This is comparatively lower than the 30-day mortality rate of 18.2% observed in patients older than 75 years from trials of fibrinolytic therapy.30 Nevertheless, in the contemporary era of PPCI, older STEMI patients remain at high risk for postinfarction complications and cardiovascular events immediately and after hospital discharge, despite provision of mechanical reperfusion.
Primary percutaneous coronary intervention using intracoronary stenting instead of balloon angioplasty only is considered the preferred reperfusion strategy in STEMI because it results in lower rates of additional revascularization procedures and potentially lower rates of MI.8,28 Relatively little is known regarding the characteristics and outcome of older patients treated with PPCI for STEMI, particularly because previous clinical trials have had low representation of older patients and/or have not reported complications and outcomes specifically by age.31- 33 In the Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications (CADILLAC) study,32 patients with acute MI, defined as STEMI, left bundle branch block, or angiographically high-grade stenosis with regional wall motion abnormalities, were randomized to angioplasty or stenting with or without abciximab. Among the cohort of 273 patients older than 75 years from CADILLAC, the 30-day and 1-year mortality rates were 4.8% and 11.1%, respectively. Guagliumi et al22 specifically evaluated elderly patients (65 years or older, n = 743) from CADILLAC and identified no differences in the rates of postprocedural flow grades by age. These findings differ from our study because we demonstrated a 2-fold higher 30-day mortality rate in patients 75 years or older (10.8%) despite PPCI with high rates of intracoronary stenting and significantly lower postprocedural TIMI grade 3 flow compared with younger patients. In addition, postinfarction event rates related to CHF or shock, which occur more often in older patients, were not reported according to age from CADILLAC.22,32
Our findings have some notable differences from those of the Primary Angioplasty in Myocardial Infarction (PAMI) trials,31,34,35 which enrolled patients between 1990 and 1998, an era predating routine use of thienopyridine, glycoprotein IIb/IIIa inhibitors, and intracoronary stenting for STEMI. Using the PAMI databases, DeGeare et al21 described characteristics, in-hospital events, and outcomes in patients 75 years or older (n = 452) compared with those younger than 75 years (n = 2580). Older patients had greater comorbidities at baseline, lower angiographic procedural success rates, and higher rates of postinfarction complications. The rate of stroke or transient ischemic attack was 2.9% in patients 75 years or older and similar to our in-hospital stroke incidence rate of 2.8%. In-hospital mortality was 10.2% in patients 75 years or older, findings similar to 30-day mortality rates from our study (10.8% in patients ≥75 years old). The PAMI investigators also identified age as a strong predictor of in-hospital mortality by entering age of 75 years or older as a categorical variable (odds ratio, 5.9; 95% confidence interval, 2.9-12.2).21 Our results, with twice the number of patients 75 years or older in the contemporary era, confirm the prognostic significance of age in patients treated with PPCI on 90-day mortality. Therefore, when taken together, our study and that performed by DeGeare et al21 suggest that despite the progress made in pharmacoinvasive management of STEMI during the past decade, older STEMI patients undergoing PPCI remain at high risk for postinfarction complications and mortality.
Prior studies comparing fibrinolysis with PPCI have demonstrated that older age is associated with adverse outcomes regardless of reperfusion strategy,10 although benefits in early outcomes and reduction in recurrent ischemia with PPCI have been suggested.36 Holmes et al10 analyzed patients from the angiographic substudy of the Global Use of Strategies to Open Occluded Coronary Arteries IIb study37 and identified a log-linear relationship between age and outcome, which remained similar regardless of whether patients were treated with fibrinolysis or primary angioplasty. Indeed, even when timely reperfusion is administered in older patients with STEMI, the risks for adverse cardiovascular events and non–cardiac-related morbidity and mortality remain high.2 Poorer outcomes in older STEMI patients may be attributable to greater comorbidities, hemodynamic status at presentation, severity of coronary artery disease, postprocedural angiographic results, and risks for postinfarction complications. The possibility of unmeasured characteristics beyond chronological age only in older patients, such as homeostatic reserve, frailty, and preexisting disability, should also be considered.38,39
In our study, patients 75 years or older (17.0%) accounted for a larger proportion of participants than comparable populations: 14% in the pooled Virtual Coordinating Centre for Global Collaborative Cardiovascular Research fibrinolytic trials and 13% in CADILLAC.2,22 Previous pooled studies of PPCI vs fibrinolysis have suggested greater absolute reduction in risk of mortality and subsequent infarction, and fewer hemorrhagic complications in all patients with PPCI; however,2,40 efforts to compare these strategies in older patients remain challenging.41 A pooled meta-analysis of patient-level data from 22 randomized trials of PPCI vs fibrinolysis demonstrated a beneficial treatment effect of PPCI over fibrinolysis for reduction of death, subsequent infarction, and stroke with no evidence of an interaction between age and treatment.42 Although evidence favoring PPCI over fibrinolysis was suggested in patients older than 80 years (P = .049), including greater absolute benefits for PPCI in elderly patients, the small sample of PPCI-treated patients in this age group (200 patients >80 years old) limited definitive conclusions regarding benefits of PPCI over fibrinolysis.
Our analysis is observational and may be influenced by unmeasured confounders. Outcomes beyond 90-day events were not available for this analysis. We did not perform an external validation of our model results in an independent population; however, internal validation was performed using bootstrapping. Patient selection in APEX-AMI was left to the discretion of the investigator; thus, older patients with serious medical conditions may not have been included in the trial. Information regarding the population screened for APEX-AMI was not available and potentially limited our ability to evaluate to what extent selection bias may have been present in this study. However, as we have described, APEX-AMI had fewer exclusion criteria than most trials of acute MI, and patients in shock or with renal failure were not excluded from participation. These features add to the generalizability of the study's findings. Patients with altered mental status or confusional states, conditions that are more common in elderly patients during times of acute illness, may not have been represented in APEX-AMI because of an inability to obtain informed consent. Despite this, we observed more comorbidities, lower angiographic success, higher complication rates, and adverse cardiac events in the oldest age strata. Nevertheless, bias due to case selection likely underestimates the risk of STEMI patients treated with PPCI in the community.
Despite optimal mechanical reperfusion and high rates of adjunctive medical treatment, age remains the main predictor of 90-day mortality in STEMI patients treated with PPCI. Efforts to attenuate this risk and understand reperfusion factors that increase age-associated outcomes are needed.
Correspondence: Christopher B. Granger, MD, Division of Cardiology, Department of Medicine, Duke Clinical Research Institute, Duke University Medical Center, 2400 Pratt St, Room 0311 Terrace Level, Durham, NC 27707 (firstname.lastname@example.org).
Accepted for Publication: September 28, 2010.
Author Contributions: All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Gharacholou, Lopes, Mehta, Pieper, Armstrong, and Granger. Acquisition of data: Gharacholou, James, Armstrong, and Granger. Analysis and interpretation of data: Lopes, Alexander, Stebbins, Pieper, James, Armstrong, and Granger. Drafting of the manuscript: Gharacholou, Lopes, and Granger. Critical revision of the manuscript for important intellectual content: Gharacholou, Lopes, Alexander, Mehta, Stebbins, Pieper, James, and Armstrong. Statistical analysis: Mehta, Stebbins, and Pieper. Obtained funding: Armstrong and Granger. Study supervision: Lopes, Alexander, James, Armstrong, and Granger.
Financial Disclosure: None reported.
Funding/Support: Dr Gharacholou is a participant in the National Institutes of Health Clinical Research Loan Repayment Program (1L30 AG034828-01).