Search tips
Search criteria 


Logo of thijTexas Heart Institute JournalSee also Cardiovascular Diseases Journal in PMCSubscribeSubmissionsTHI Journal Website
Tex Heart Inst J. 2009; 36(1): 24–30.
PMCID: PMC2676527

Clinical Predictors of Late Death in Survivors of Acute Myocardial Infarction


Survivors of acute myocardial infarction have higher mortality rates than do the general population. This study examined the value of multiple clinical characteristics in predicting late death among patients who present with acute myocardial infarction.

We reviewed the electronic medical records of patients who had been treated for acute myocardial infarction at our institution from 1992 through 2000. We abstracted the clinical, laboratory, electrocardiographic, echocardiographic, and treatment characteristics.

Of 144 patients (79.2% men; 97.2% white; mean age, 63 ± 14.2 yr) included in this analysis, 63 (43.8%) patients died during a follow-up period of 5.6 ± 2.8 years (5 d–12.7 yr). Higher age (hazard ratio, 1.83 ± 0.31 for every 10-year increase), elevated serum creatinine (hazard ratio, 2.87 ± 0.76), and lower baseline left ventricular ejection fraction (hazard ratio, 0.74 ± 0.21 for every 5% increase) were found to be predictors of late death after adjusting for the white blood cell count, the QRS duration, the presence of coronary revascularization or defibrillator implantation, and the history of coronary artery disease. Elevated white blood cell count predicted early but not late death. Patients with none of the above risk factors had 100% survival at 5 years, in comparison with 22.7% survival for those with 3 or more of the 4 risk factors identified above.

In this study, we have identified clinical predictors of long-term survival after acute myocardial infarction that might help in prognostication, patient education, and risk modification.

Key words: Acute disease, European continental ancestry group, mortality, late, myocardial infarction/mortality, predictive value of tests, prognosis, retrospective studies, risk assessment, survival analysis

Acute myocardial infarction (AMI) is experienced by more than 800,000 people in the United States annually. Although approximately 75% of AMI victims survive beyond the 1st year, they have significantly higher mortality rates than does the general population.1

Several studies have contributed to our understanding of characteristics at presentation that predict death. Most of them, however, have focused on predictors of death at 30 days or less,2–5 or during the 1st year,6–11 and very few have examined long-term mortality rates.12 Furthermore, most of the above studies have examined only a few clinical factors simultaneously. Some of the predictors of death were higher age, larger infarct size, elevated cardiac tissue-injury markers, impaired renal function, elevated blood-sugar levels, elevated white blood cell (WBC) counts, decreased left ventricular ejection fraction (LVEF), and increased QRS duration.

This current study examined retrospectively the prognostic value of various clinical characteristics on late, all-cause mortality rates in patients presenting with AMI at a single, large, tertiary center.


We reviewed a database of patients who had presented with AMI with and without concurrent cardiac arrest at the University of Pittsburgh Medical Center from 1992 through 2000. The database was originally built by including all patients who presented with concurrent AMI and cardiac arrest, who were then matched by age, sex, and left ventricular systolic function in a 2:1 ratio to a group of patients who presented with AMI only. It is worth noting that this preexisting database, which originally had been designed to compare the clinical characteristics of patients presenting with AMI and cardiac arrest to those of patients presenting with AMI only, did not include all AMI patients seen at our institution during the study period.

The current study was approved by the Institutional Review Board of the University of Pittsburgh, who waived the need for informed consent for enrolled patients. Acute myocardial infarction was defined in accordance with the revised criteria13 of the World Health Organization as the concurrence of 2 of the following 3 findings: chest pain; ST-T segment changes on surface electrocardiography; and elevation in cardiac enzymes (creatine kinase–MB fraction or troponin, or both). Abstracted data included baseline demographic characteristics, comorbid conditions, electrocardiographic (ECG) data, LVEF, laboratory data including cardiac enzymes collected within 6 hours of admission, records of current and previous revascularization procedures, presence of an implantable cardioverter-defibrillator, and medications prescribed at hospital discharge. Patients' mortality status was ascertained from the electronic medical records, the Social Security Mortality Index (, or both. Patients who were not classified as dead by the above 2 methods were censused on 30 November 2006. For the purpose of this study, early and late death were defined, respectively, as death within or beyond 2 years of the occurrence of AMI.

Statistical Analysis

Means and percentages were used to describe continuous and categorical variables, respectively. In order to identify predictive variables, we first performed univariate analyses by using the log-rank test and the Cox proportional hazard regression model for categorical and continuous variables, which included matched variables. Further, multivariate Cox proportional hazard regression models were used to select variables with an acceptable alpha error ≤0.1, using backward elimination. The variables selected by the above model, along with additional relevant variables from the medical literature,8,9,12 were used to build a final model after checking for interactions and proportionality assumptions. The proportionality hazard assumption was evaluated by including continuous-time interaction terms and the Cox test, along with visual inspection of Kaplan-Meier curves. Kaplan-Meier curves were plotted for variables that had been found significant on the multivariate Cox model by dichotomizing markers according to clinical relevance. Finally, the effect of the presence of 1 or more of the dichotomized risk predictors (age, creatinine level, QRS duration, WBC count) was evaluated by plotting Kaplan-Meier survival curves and by using the log-rank test. P values <0.05 were considered statistically significant. We used SAS version 9.1.3 (SAS Institute; Cary, NC) for all statistical analyses.


We reviewed the electronic medical records of 144 patients (79.2% men; 97.2% white; mean age, 63 ± 14.2 yr) who had presented with AMI at the University of Pittsburgh Medical Center. Of those, 63 (43.8%) patients died during a follow-up period of 5.6 ± 2.8 years (range, 5 d–12.7 yr). At presentation, ST-segment elevation myocardial infarction (STEMI) had been present in 36% of patients, congestive heart failure (CHF) in 14.6%, and concurrent near-fatal arrhythmia in 33%. The survival rates at 1, 5, and 10 years after presentation with AMI were 90.3%, 66.7%, and 42.4%, respectively. Table I shows the results of univariate analysis for abstracted patient characteristics at the time of presentation with AMI. The following characteristics were found to be associated with higher mortality rates: higher age at presentation, history of coronary artery disease (CAD), CHF, cardiomyopathy, diabetes mellitus, hypertension, chronic obstructive pulmonary disease, family history of CAD, higher serum creatinine level, longer QRS duration, atrial fibrillation, right or left bundle branch block, lower LVEF, and wall-motion hypokinesia. Many of the above characteristics were statistically correlated with one another due to their association at the pathophysiologic level or to their tendency to measure the same disease process. An elevated WBC count was found to predict early but not late death (Fig. 1).

figure 6FF1
Fig. 1 Kaplan-Meier survival curves stratified by white blood cell count (WBC; dotted line, ≥10,600/mm3). A) Early follow-up: Note that the stratified curves remain separated and were found to be statistically different at 1 year (log-rank P < ...
Table thumbnail
TABLE I. Characteristics of Patients Who Presented with Acute Myocardial Infarction, Arranged by Mortality Status during Follow-Up (n = 144)

The multivariate model with backward selection method (using all the characteristics with univariate P values <0.1) selected higher age, higher serum creatinine level (>1.3 mg/dL), and lower baseline LVEF as predictors of late death. Table II shows the hazard ratio derived from a multivariate model with use of characteristics selected by the above model, along with additional characteristics that have been shown in the recent literature8,9,12 to be predictors of total mortality rate after AMI. Higher age (hazard ratio [HR] = 1.83 ± 0.31 for every 10-yr increase), elevated serum creatinine (HR = 2.87 ± 0.76), and lower baseline LVEF (HR = 0.74 ± 0.21 for every 5% increase) were found to be predictors of late death after adjustment for the presence of a previous history of CAD, elevated WBC, wider QRS duration, revascularization, or defibrillator implantation. In our data, there was no statistical interaction between any of these independent predictors of time to death. Also, the Kaplan-Meier curves or Cox tests for proportional hazard assumption were not significant either collectively or for any of the final variables in the model.

Table thumbnail
TABLE II. Multivariate Cox Proportional Hazard-Ratio Model

Finally, we created disjointed categories by dichotomizing the 3 variables that had been found significant in the above model, and to these we added WBC count, due to its importance in the medical literature; accordingly, we created a score of 0, 1, 2, or 3+ to indicate the presence of zero, 1, 2, or “3 or more” risk factors. Age was dichotomized at the population mean age of 63 years, QRS duration at 120 ms, serum creatinine level at 1.3 mg/dL, and WBC count at 10,600/mm3—all values that represent the upper limits of normal for adults at our institution. Kaplan-Meier plots (Fig. 2) demonstrate that patients with no risk factors had a 100% 5-year survival, compared with 82.4%, 60.4%, and 22.7% in patients with 1, 2, and 3 or more risk factors, respectively (P < 0.001).

figure 6FF2
Fig. 2 Kaplan-Meier graph showing the influence of the presence of 3 or more (n = 22), 2 (n = 48), 1 (n = 51), or none (n = 14) of the 4 risk factors (age ≥63 yr; left ventricular ejection fraction <0.30; ...


This study has several limitations. First, it is a retrospective, single-center study, with a small sample size that consists mostly of white men. It mixes patients who had STEMI and non-STEMI, including those who had cardiac arrest concomitant with the AMI. Although all-cause mortality rate is a very important cardiovascular endpoint, disease-specific mortality data would have provided valuable incremental information; unfortunately, these data were not available. Last, the present study spans more than a decade. It is therefore conceivable that changes in the standards of clinical management of AMI patients throughout this period influenced the results presented herein.


In contrast with most studies that examine predictors of early death after AMI, our data enabled us to examine late death (mean follow-up, 5.6 yr) and found age, higher serum creatinine levels, and lower baseline LVEF to be important predictors of death for patients who presented with AMI. Interestingly, elevated WBC count was found to be a predictor of early death but not of late death. Identification of these factors could lead to better prognostication and hence to improved patient counseling. Also, there could be an attempt to modify the pathologic process that underlies some of these characteristics, which could result in better outcomes after AMI.

Elderly patients are at a higher risk of AMI and have higher rates of AMI-related complications and cardiovascular death. Early rehabilitation through incremental-walking and other post-discharge exercise programs, and education of the patient and family, are important issues that require special attention.14 Whether any of these interventions can influence outcomes in the elderly after AMI remains to be shown. Moreover, a study that made use of nationwide data15 reported that the elderly population receives lower use of coronary artery bypass grafting than does the younger population; extending coronary reperfusion therapy to the elderly might also improve the poor outcomes seen in older patients.16

Renal function impairment as indicated by higher serum creatinine levels has been shown to be a strong predictor of death in studies that examine all AMI patients,5,6,10,11,17 both STEMI3,9 and non-STEMI.7 There is tremendous increase (>20 times) in the rate of cardiovascular death in patients with end-stage renal disease, compared with the general population.18 However, even mild renal impairment is an important marker of increased mortality risk, as shown by the above studies and as confirmed by our current analysis. This suggests that early management of renal dysfunction is important not only to avoid advanced or end-stage kidney disease but also to possibly reduce the risk of cardiovascular death.

A higher WBC count early after AMI may be a marker of concomitant infection; however, it also correlates well with infarct size and with the extent of inflammatory response after AMI.19 After an ST-segment elevation AMI, reperfusion appears to lower WBC count.20 Similarly, the use of anti-inflammatory drugs such as aspirin21 and statins22 lowers the WBC count and appears to improve survival. Elevated WBC count in our current study was found to predict death only up to 1 or 2 years after AMI (Fig. 1A), but not late death (Fig. 1B). Most studies that have reported a strong association between elevated WBC count and death have focused on in-hospital mortality rates or on mortality rates early after discharge—hence, the association needs more scrutiny. It is important to note that there is an increase in WBC count about 3 to 5 days after the acute presentation of myocardial infarction.14 White blood cell count during this period may be a marker of the extent of myocardial injury and therefore may correlate with diminished LVEF and with elevated late mortality rate. Although elevated neutrophil counts (in comparison with elevated counts of other WBC subtypes23) are associated with poorer prognoses, the role of such elevation in predicting late death needs further exploration.

Low LVEF values were a significant predictor of late death in our data set. Other studies have reported similar findings after adjusting for age, ST-segment elevation, cardiac enzymes, and TIMI risk score.8 Although an implantable cardioverter-defibrillator may save a life in the event of near-fatal arrhythmia, it may not help in the event of decompensated cardiac function. In this circumstance, pharmacologic interventions with agents such as angiotensin-converting enzyme inhibitors and β-blockers play an important role in management, through structural remodeling and other mechanisms.24

Increased QRS duration after AMI has also been associated with early and late cardiac death.25–27 Its prognostic role is seen especially in patients who have non-STEMI. Mechanically, the higher mortality rate conferred by increased QRS duration may be explained by its association with CHF, by a higher atherosclerotic burden, and by life-threatening arrhythmias.27

Although diabetes was a significant predictor of death after AMI in our univariate model, it was not significant in our multivariate model, which could be the result of limited power. We have not abstracted the stress or fasting blood-glucose levels. Other studies have shown that higher blood-glucose levels among non-diabetic patients presenting with AMI, followed by almost stable repeat glucose levels within 24 hours of presentation, are a strong predictor of death.28 Interestingly, 1 study has shown a lower mortality rate among black AMI patients with diabetes than among white AMI patients with diabetes.15 These issues need further clarification.

One study4 has reported an additive effect of multiple risk factors, including decreased LVEF and impaired renal function, on total mortality after AMI. In our analysis, we explored the combined effect of older age, renal impairment, decreased LVEF, and elevated WBC count. Interestingly, the 5-year survival rate was 100% among patients with no risk factor, in comparison with only 22.7% among those with 3 or more risk factors. It would be valuable to have late-mortality risk models that have been derived and validated through use of existing data sets with larger sample sizes.


The multiple risk factors presented in our study strongly predict late death after AMI and could be of value in prognostication and patient education. Also, some risk factors identified here, such as inflammation and low LVEF, can be modified via such interventions as revascularization, the use of angiotensin-converting enzyme inhibitors and β-blockers, and the use of proven anti-inflammatory drugs. Future studies that examine the possible effects of these interventions on late death after AMI are needed.


Address for reprints: Samir Saba, MD, Cardiac Electrophysiology Section, University of Pittsburgh Medical Center, 200 Lothrop St., PUH B535, Pittsburgh, PA 15213. E-mail: ude.cmpu@sabas


1. Ryan TJ, Antman EM, Brooks NH, Califf RM, Hillis LD, Hiratzka LF, et al. 1999 update: ACC/AHA guidelines for the management of patients with acute myocardial infarction. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Management of Acute Myocardial Infarction). J Am Coll Cardiol 1999;34(3):890–911. [PubMed]
2. Carasso S, Sandach A, Beinart R, Schwammenthal E, Sagie A, Kuperstein R, et al. Usefulness of four echocardiographic risk assessments in predicting 30-day outcome in acute myocardial infarction. Am J Cardiol 2005;96(1):25–30. [PubMed]
3. Timoteo AT, Fiarresga A, Feliciano J, Pelicano N, Ferreira L, Ferreira R, et al. The prognostic impact of renal failure in patients with ST-segment elevation acute myocardial infarction [in English, Polish]. Kardiol Pol 2005;63(4):373–80. [PubMed]
4. Marenzi G, Moltrasio M, Assanelli E, Lauri G, Marana I, Grazi M, et al. Impact of cardiac and renal dysfunction on inhospital morbidity and mortality of patients with acute myocardial infarction undergoing primary angioplasty. Am Heart J 2007;153(5):755–62. [PubMed]
5. Schiele F, Seronde MF, Descotes-Genon V, Blonde MC, Legalery P, Meneveau N, et al. Impact of renal dysfunction and glucometabolic status on one month mortality after acute myocardial infarction. Acute Card Care 2007;9(1):34–42. [PubMed]
6. Masoudi FA, Plomondon ME, Magid DJ, Sales A, Rumsfeld JS. Renal insufficiency and mortality from acute coronary syndromes. Am Heart J 2004;147(4):623–9. [PubMed]
7. Gibson CM, Dumaine RL, Gelfand EV, Murphy SA, Morrow DA, Wiviott SD, et al. Association of glomerular filtration rate on presentation with subsequent mortality in non-ST-segment elevation acute coronary syndrome; observations in 13,307 patients in five TIMI trials. Eur Heart J 2004;25 (22): 1998–2005. [PubMed]
8. Bosch X, Theroux P. Left ventricular ejection fraction to predict early mortality in patients with non-ST-segment elevation acute coronary syndromes. Am Heart J 2005;150(2):215–20. [PubMed]
9. Goldberg A, Hammerman H, Petcherski S, Zdorovyak A, Yalonetsky S, Kapeliovich M, et al. Inhospital and 1-year mortality of patients who develop worsening renal function following acute ST-elevation myocardial infarction. Am Heart J 2005;150(2):330–7. [PubMed]
10. Schiele F. Chronic renal failure: an independent factor of mortality after myocardial infarction [in French]. Ann Cardiol Angeiol (Paris) 2005;54(4):161–7. [PubMed]
11. Yan AT, Yan RT, Tan M, Constance C, Lauzon C, Zaltzman J, et al. Treatment and one-year outcome of patients with renal dysfunction across the broad spectrum of acute coronary syndromes. Can J Cardiol 2006;22(2):115–20. [PMC free article] [PubMed]
12. Nordin C, Amiruddin R, Rucker L, Choi J, Kohli A, Marantz PR. Diabetes and stress hyperglycemia associated with myocardial infarctions at an urban municipal hospital: prevalence and effect on mortality. Cardiol Rev 2005;13(5):223–30. [PubMed]
13. Nomenclature and criteria for diagnosis of ischemic heart disease. Report of the Joint International Society and Federation of Cardiology/World Health Organization task force on standardization of clinical nomenclature. Circulation 1979; 59(3):607–9. [PubMed]
14. Graham I, Ingram S, Fallon N, Leong T, Gormley J, O'Doherty V, et al. Rehabilitation of the patient with coronary heart disease. In: Fuster V, O'Rourke RA, Walsh RA, Poole-Wilson P, editors. Hurst's the heart. 12th ed. New York: McGraw-Hill Medical Publishing Division, 2008.
15. Kamalesh M, Subramanian U, Ariana A, Sawada S, Peterson E. Diabetes status and racial differences in post-myocardial infarction mortality. Am Heart J 2005;150(5):912–9. [PubMed]
16. Balzi D, Barchielli A, Buiatti E, Franceschini C, Mangani I, Del Bianco L, et al. Age and comorbidity in acute myocardial infarction: a report from the AMI-Florence Italian registry. Am J Geriatr Cardiol 2006;15 (1):35–41. [PubMed]
17. Schiele F, Legalery P, Didier K, Meneveau N, Seronde MF, Caulfield F, et al. Impact of renal dysfunction on 1-year mortality after acute myocardial infarction. Am Heart J 2006;151 (3):661–7. [PubMed]
18. Foley RN, Parfrey PS, Sarnak MJ. Clinical epidemiology of cardiovascular disease in chronic renal disease. Am J Kidney Dis 1998;32(5 Suppl 3):S112-9. [PubMed]
19. Mulvihill NT, Boccalatte M, Foley JB. Inflammatory markers as predictors of clinical outcome in acute coronary syndromes. Minerva Cardioangiol 2002;50(6):653–9. [PubMed]
20. Smit JJ, Ottervanger JP, Slingerland RJ, Suryapranata H, Hoorntje JC, Dambrink JH, et al. Successful reperfusion for acute ST elevation myocardial infarction is associated with a decrease in WBC count. J Lab Clin Med 2006;147(6):321–6. [PubMed]
21. Becker RC. Antiplatelet therapy in coronary heart disease. Emerging strategies for the treatment and prevention of acute myocardial infarction. Arch Pathol Lab Med 1993;117(1): 89–96. [PubMed]
22. Ray KK, Cannon CP, Ganz P. Beyond lipid lowering: What have we learned about the benefits of statins from the acute coronary syndromes trials? Am J Cardiol 2006;98(11A):18P-25P. [PubMed]
23. Dragu R, Huri S, Zuckerman R, Suleiman M, Mutlak D, Agmon Y, et al. Predictive value of white blood cell subtypes for long-term outcome following myocardial infarction. Atherosclerosis 2008;196(1):405–12. [PubMed]
24. Ambrosioni E, Borghi C, Magnani B. The effect of the angiotensin-converting-enzyme inhibitor zofenopril on mortality and morbidity after anterior myocardial infarction. The Survival of Myocardial Infarction Long-Term Evaluation (SMILE) Study Investigators. N Engl J Med 1995;332(2): 80–5. [PubMed]
25. Korhonen P, Husa T, Tierala I, Vaananen H, Makijarvi M, Katila T, et al. QRS duration in high-resolution methods and standard ECG in risk assessment after first and recurrent myocardial infarctions. Pacing Clin Electrophysiol 2006;29(8): 830–6. [PubMed]
26. Hathaway WR, Peterson ED, Wagner GS, Granger CB, Zabel KM, Pieper KS, et al. Prognostic significance of the initial electrocardiogram in patients with acute myocardial infarction. GUSTO-I Investigators. Global Utilization of Streptokinase and t-PA for Occluded Coronary Arteries. JAMA 1998;279(5):387–91. [PubMed]
27. Brilakis ES, Mavrogiorgos NC, Kopecky SL, Rihal CC, Gersh BJ, Williams BA, Clements IP. Usefulness of QRS duration in the absence of bundle branch block as an early predictor of survival in non-ST elevation acute myocardial infarction. Am J Cardiol 2002;89(9):1013–8. [PubMed]
28. Goyal A, Mahaffey KW, Garg J, Nicolau JC, Hochman JS, Weaver WD, et al. Prognostic significance of the change in glucose level in the first 24 h after acute myocardial infarction: results from the CARDINAL study. Eur Heart J 2006; 27(11):1289–97. [PubMed]

Articles from Texas Heart Institute Journal are provided here courtesy of Texas Heart Institute