The causes of death within 1 year of hospital admission in patients with non–ST-segment elevation acute coronary syndromes are ill defined, particularly in patients aged ≥75 years. From January 2008 through May 2010, we enrolled 645 patients aged ≥75 years with non–ST-segment elevation acute coronary syndromes: 313 in a randomized trial comparing an early aggressive versus an initially conservative approach, and 332, excluded from the trial for specific reasons, in a parallel registry. Each death occurring during 1 year of follow-up was adjudicated by an independent committee. The mean age was 82 years in both study cohorts, and 53% were men. By the end of the follow-up period (median 369 days, interquartile range 345 to 391), 120 patients (18.6%) had died. The mortality was significantly greater in the registry (23.8% vs 13.1%, p = 0.001). The deaths were classified as cardiac in 94% of the cases during the index admission and 68% of the cases during the follow-up period. Eighty-six percent of the cardiac deaths were of ischemic origin. In a multivariate logistic regression model that included the variables present on admission in the whole study population, the ejection fraction (hazard ratio 0.95, 95% confidence interval 0.94 to 0.97; p <0.001), hemoglobin level (hazard ratio 0.85, 95% confidence interval 0.76 to 0.94; p = 0.001), older age (hazard ratio 1.05, 95% confidence interval 1.01 to 1.10, p = 0.010), and creatinine clearance (hazard ratio 0.99, 95% confidence interval 0.97 to 0.99; p = 0.030) were the independent predictors of all-cause death at 1 year. In conclusion, within 1 year after admission for non–ST-segment elevation acute coronary syndromes, most deaths in patients aged ≥75 years have a cardiac origin, mostly owing to myocardial ischemia.
No study has specifically investigated the causes of death (CoD) in patients with non–ST-segment elevation acute coronary syndrome (NSTEACS), particularly in those aged ≥75 years, who represent about 30% of the patients admitted to the hospital with an NSTEACS diagnosis. The Italian Elderly ACS study, the first study specifically investigating treatment strategies for elderly patients with NSTEACS, represented a unique opportunity to investigate this issue. At 23 participating hospitals, the study enrolled 645 patients with an average age of 82 years: 313 patients were included in a randomized clinical trial (RCT) of an early invasive strategy compared with an initially conservative strategy and 332 in a parallel registry of those excluded from the RCT for any reason. The results of the RCT showed no significant benefit of an early invasive approach in the whole study population, although the patients with elevated troponin levels on admission showed a significant reduction in the primary end point of death, myocardial infarction, stroke, and rehospitalization within 1 year. The purpose of the present analysis was to investigate the CoD of the patients enrolled in the whole study and to find the independent predictors of mortality.
Methods
The Italian Elderly ACS study enrolled patients aged ≥75 years with NSTEACS who were admitted within 48 hours of the most recent ischemic symptoms and showed either ischemic electrocardiographic changes or elevated cardiac markers, or both. The clinical trial registration ( ClinicalTrials.gov ) number was NCT00510185 . The main exclusion criteria were secondary causes of myocardial ischemia, ongoing myocardial ischemia or acute heart failure despite maximally titrated treatment, percutaneous coronary intervention or coronary bypass surgery within 30 days before randomization, a serum creatinine level >2.5 mg/dl, a history of cerebrovascular accident within the previous month, active internal bleeding or recent transfusions, gastrointestinal or genitourinary bleeding within 6 weeks before randomization, a platelet count <90,000 cells/μl, ongoing oral anticoagulation, severe obstructive lung disease, malignancy, or a neurologic deficit limiting follow-up. The patients meeting the inclusion criteria for the study, but who were excluded from the RCT for any reason, were to be enrolled in the registry. The patients enrolled in the RCT were randomized to an early aggressive strategy (i.e., coronary angiography within 72 hours and, if indicated, revascularization) or to an initially conservative strategy (with angiography or revascularization only in the case of recurrent ischemic symptoms). The patients included in the registry were treated according to hospital routine in the specific cases. After discharge, follow-up visits were planned at 1, 6, and 12 months, with a maximum allowance to 405 days.
The clinical data and relevant clinical events were collected using an electronic case record form and were subjected to source data verification, with special emphasis on the events forming the combined primary end point of death from any cause, myocardial infarction, stroke, and rehospitalization for cardiovascular causes or severe bleeding. In both cohorts of the study, to be eligible, the patients had to be able to sign an informed consent form. The ethics committees at each participating center approved the present study.
The CoDs were divided into cardiovascular, secondary to bleeding, and noncardiovascular. The cardiovascular CoDs were further divided according to 2 main categories: cardiac (subcategorized as ischemic heart disease, cardiogenic shock, and arrhythmias) and vascular (subcategorized as stroke, pulmonary embolism, and peripheral artery disease). Ischemic death was defined according to the MONItoring trends and determinants of CArdiovascular disease (MONICA) criteria. Thus, the underlying disease was further divided into (1) definite myocardial infarction; (2) possible acute myocardial infarction or coronary death; and (3) fatal cases with insufficient data. The cases of myocardial ischemia secondary to stent thrombosis were considered a separate category among the ischemic CoDs. Only definite and probable stent thrombosis, according to the definition proposed by the Academic Research Consortium, were included. Cardiogenic shock was considered the primary CoD when persistent hypotension and end organ failure or pulmonary edema were present in association with severe left ventricular dysfunction or hemodynamic signs of severe right ventricular infarction in the absence of ongoing cardiac ischemia or other extracardiac etiologies. Arrhythmic death was defined, using the Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto (GISSI)-3 definition, as the abrupt spontaneous cessation of respiration and blood circulation when ventricular tachycardia, ventricular fibrillation, or asystole were recorded within 10 minutes of clinical death, in the absence of signs or symptoms of ventricular failure or pericardial tamponade. Arrhythmias clearly due to ongoing myocardial ischemia were excluded. Vascular deaths included all deaths secondary to pulmonary embolism, defined using the appropriateness of diagnostic management applied by the Etude de pratique des Méthodes Diagnostiques d’Embolie Pulmonaire aux Urgences (EMDEPU) study group ; stroke, defined as any new neurologic deficit lasting >24 hours and confirmed by a neurologist report or computed tomography or magnetic resonance imaging findings to distinguish between ischemic and hemorrhagic stroke; any complication of peripheral embolization (e.g., intestinal, leg, or renal), as documented by the clinical records; and dissecting aneurysm, as documented by computed tomography, magnetic resonance imaging, or autopsy. Deaths secondary to bleeding were considered those occurring as a direct consequence of major bleeding, defined according to the Global Utilization of Streptokinase and t-PA [tissue plasminogen activator] for Occluded Coronary Arteries (GUSTO) criteria and in accordance with the Academic Bleeding Consensus Multidisciplinary Working Group recommendations. Only intracranial hemorrhage and bleeding leading to hemodynamic deterioration were considered a direct CoD. Noncardiovascular death was considered in the absence of any of the above causes of death and in the presence of extracardiovascular etiologies. The main underlying disease was divided into respiratory, metabolic, cancer, trauma, and renal insufficiency. For noncardiovascular death, the Swedish Causes of Death Register classification was adopted. Accordingly, respiratory system-specific mortality was subdivided into chronic obstructive pulmonary disease, influenza, and pneumonia. Metabolic was subdivided into diabetic and nondiabetic. Death from cancer was subdivided into lung, colorectal, stomach, pancreatic, breast, and prostate cancer. Finally, “others” was considered death from trauma, including complications due to any accidental trauma. A separate category was considered for death from renal insufficiency.
The CoDs were adjudicated by an independent clinical event committee. If death occurred during hospitalization (either during the index admission or during subsequent hospitalizations), the participant’s hospital records, including the medical history, physical examination findings, laboratory findings, autopsy findings, and discharge diagnosis, were abstracted by the study investigators using a standard form. In addition, photocopies of selected sections of the participant’s inpatient record, discharge summary, electrocardiogram, and pathology reports were obtained. In the cases of deaths occurring out of the hospital, efforts were made to obtain the death certificate. If this was not possible, the verbal autopsy method was applied. This method uses information obtained from a close relative or caretaker of a deceased person about the circumstances, symptoms, and signs during the terminal illness to assign a CoD.
The primary analysis of the present study concerned all deaths occurring within 405 days of follow-up. Descriptive statistics were computed for all potential covariates. Numbers and percentages are used to describe categorical variables and median and interquartile range to describe continuous variables. Categorical variables were compared using the chi-square test and continuous variables using the Student t test. The 1-year survival was estimated using the Kaplan-Meier method. Univariate and multivariate Cox regression analyses have been used to identify variables independently associated with mortality. The proportional hazard assumption was verified using Shoenfeld residuals. A p value <0.05 was considered statistically significant. Two separate all-cause mortality models were created for the trial and the registry, with a final model for the overall study population that included only the variables equally distributed between the trial and registry and identified as significant on univariate analysis. In addition, the total number of all-cause, cardiovascular, and noncardiovascular deaths during the study period within both populations were determined. Finally, the model was redone for the univariate and multivariate analyses, with the total population in mind to explore cardiovascular mortality. We assessed for evidence of effect modification among pairs of predictors, but we found no substantive interactions. We did not include interaction terms in the model. The estimates of the association between the predictors and end points are presented as hazard ratios and 95% confidence intervals (CIs). The Wald chi-square for all factors in the models are summed. The percentage for each chi-square value relative to this total was used as an estimate of the percentage of the predictive information in the model accounted for by that variable. Analysis was performed using the Statistical Package for Social Sciences, versions 15.0 and 19 (SPSS, Chicago, Illinois).
Results
A total of 645 patients agreed to participate in the present study. Of these, 313 were enrolled in the RCT and 332 in the registry. The baseline characteristics of the patients are reported in Table 1 . The patients included in the registry had significantly worse renal function and ejection fraction and lower blood hemoglobin levels. Also, elevated troponin levels on admission and ischemic electrocardiographic changes were significantly more prevalent in the registry population. Coronary angiography and myocardial revascularization procedures during the index admission were performed equally in the 2 cohorts; however, the patients in the registry received significantly less guideline-recommended drug therapy, both during the index admission and at discharge ( Table 2 ). A total of 11 patients (1.7%) were lost to follow-up, 4 in the RCT and 7 in the registry.
| Variable | RCT (n = 313) | Registry (n = 332) | Overall Study Population (n = 645) | p Value |
|---|---|---|---|---|
| Age (yrs) | 81 (77–84) | 82 (78–85) | 81 (78–85) | 0.072 |
| Men | 157 (50) | 187 (56) | 344 (53) | 0.117 |
| Diabetes mellitus | 114 (36) | 117 (30) | 231 (36) | 0.755 |
| Creatinine clearance ∗ (ml/min) | 51 (40–61) | 43 (29–58) | 47.5 (35–60) | <0.0001 |
| Atrial fibrillation | 42 (13) | 45 (14) | 87 (14) | 0.397 |
| Previous stroke | 25 (7.9) | 33 (9.9) | 58 (8.9) | 0.043 |
| Elevated troponin levels | 190 (61) | 293 (88) | 483 (75) | <0.0001 |
| ECG ischemic changes | 198 (63) | 308 (93) | 506 (78) | <0.0001 |
| Ejection fraction (%) | 50 (40–55) | 48 (36–55) | 50 (40–55) | 0.008 |
| Hemoglobin (g/dl) | 13.2 (12.0–14.4) | 12.9 (11.3–14.0) | 13.1 (11.7–14.3) | 0.002 |
∗ Creatinine clearance calculated using the Cockroft-Gault formula: creatinine clearance (ml/min) = (0.85 for women or 1.00 for men) × (140 − age [years]) × body weight [kg]/72 × serum creatinine [mg/dl].
| Variable | RCT (n = 313) | Registry (n = 332) | Overall Study Population (n = 645) | p Value |
|---|---|---|---|---|
| Coronary angiography | 182 (58) | 197 (59) | 379 (59) | 0.282 |
| PCI | 111 (36) | 133 (40) | 244 (38) | 0.229 |
| CABG | 10 (3.2) | 13 (3.9) | 23 (3.5) | 0.443 |
| Medication during index admission | ||||
| Aspirin | 300 (96) | 294 (89) | 594 (92) | 0.0006 |
| Ticlopidine/clopidogrel | 293 (94) | 257 (77) | 550 (85) | <0.0001 |
| GPIIb/IIIa inhibitor | 35 (11) | 46 (14) | 81 (13) | 0.306 |
| Unfractionated heparin | 76 (24) | 116 (35) | 192 (30) | 0.003 |
| Enoxaparin | 187 (60) | 195 (59) | 382 (59) | 0.829 |
| Medication at discharge | ||||
| Aspirin | 284 (91) | 275 (83) | 559 (87) | 0.006 |
| Ticlopidine/clopidogrel | 245 (78) | 230 (69) | 475 (74) | 0.009 |
| β Blockers | 185 (59) | 195 (59) | 380 (59) | 0.600 |
| ACEI/ARB | 254 (81) | 219 (66) | 473 (73) | <0.0001 |
| Statins | 251 (80) | 220 (66) | 471 (73) | 0.0002 |
As shown in Figure 1 , 32 patients died during the index hospitalization: 14 (4.5%) of those enrolled in the RCT and 18 (5.4%) of those enrolled in the registry (p = 0.58). After discharge, another 88 patients died: 27 (8.6%) of those enrolled in the RCT and 61 (18.3%) of those enrolled in the registry (p <0.001). Overall, within 1 year after admission, 41 (13.1%) of the patients enrolled in the RCT and 79 (23.8%) of those enrolled in the registry had died (odds ratio 1.95, 95% CI 1.34 to 2.84, p = 0.001 for registry compared with RCT).
The distribution between the cardiovascular and noncardiovascular CoDs was similar in the 2 cohorts. During the index admission, 30 of 32 deaths (94%) were of cardiovascular origin (all cardiac). During follow-up, cardiovascular death was equally prevalent in both the RCT (78%) and the registry (76%), with most cases related to myocardial infarction or coronary death ( Figure 1 ). No cases of fatal bleeding occurred in the whole study.
Eighty-six patients (72% of total deaths) died during the first 6 months of follow-up, and 34 thereafter (odds ratio for second vs first semester 0.27, 95% CI 0.10 to 0.69). During the first 6 months, 87% of the deaths were of cardiovascular origin compared with 65% thereafter (OR of second vs first semester 0.74, 95% CI 0.57 to 0.96). In contrast, noncardiovascular mortality contributed more after the first 6 months (OR of second vs first semester 2.76, 95% CI 1.35 to 5.64). The cumulative risk estimates for total, cardiovascular, and noncardiovascular mortality are shown in Figure 2 .
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
