Author + information
- Revision received September 20, 2000
- Revision received January 19, 2001
- Accepted January 29, 2001
- Published online May 1, 2001.
- Mark I Furman, MD, FACCa,* (, )
- Harold L Dauerman, MD, FACCa,
- Robert J Goldberg, PhDa,
- Jorge Yarzbeski, MD, MPHa,
- Darleen Lessard, MSa and
- Joel M Gore, MD, FACCa
- ↵*Reprint requests and correspondence:
Dr. Mark I. Furman, Division of Cardiovascular Medicine, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, Massachusetts 01655
The goal of this study was to examine long-term trends in the incidence, in-hospital and long-term mortality patterns in patients with an initial non-Q-wave myocardial infarction (NQWMI) as compared with those with an initial Q-wave myocardial infarction (QWMI).
Limited data are available describing trends in the incidence and mortality from an initial QWMI and NQWMI from a multi-hospital community-wide perspective.
Our study was an observational study of 5,832 metropolitan Worcester, Massachusetts residents (1990 census = 437,000) hospitalized with validated initial acute MI in all greater Worcester hospitals during 11 annual periods between 1975 and 1997.
The incidence of QWMI progressively decreased between 1975/78 (incidence rate = 171/100,000 population) and 1997 (101/100,000 population). In contrast, the incidence of NQWMI progressively increased between 1975/78 (62/100,000 population) and 1997 (131/100,000 population). Hospital death rates were 19.5% for patients with QWMI and 12.5% for those with NQWMI. After controlling for various covariates, patients with QWMI remained at significantly increased risk for hospital mortality (adjusted odds ratio = 1.63; 95% confidence interval: 1.35, 1.97). While the hospital mortality of QWMI has progressively declined over time (1975/78 = 24%; 1997 = 14%), the in-hospital mortality for NQWMI has remained the same (1975/78 = 12%; 1997 = 12%). These trends remained after adjusting for potentially confounding prognostic factors. The multivariable adjusted two-year mortality after hospital discharge declined over time for patients with QWMI and NQWMI.
Despite impressive declines in the incidence, in-hospital and long-term mortality associated with QWMI, NQWMI is increasing in frequency and has the same in-hospital mortality now as it did 22 years ago.
Coronary heart disease is often considered as a spectrum of disease states ranging from silent myocardial ischemia on one extreme to acute myocardial infarction (AMI) at the other extreme. Unstable angina and non-Q-wave myocardial infarction (NQWMI) occupy the midpoint of this pathophysiologic spectrum (1). Rupture of an atherosclerotic coronary plaque with resultant thrombus formation is a common pathophysiological feature of the acute coronary syndromes of unstable angina, NQWMI and Q-wave MI (QWMI) (2). Despite these common features, the results of several studies suggest that the short- and long-term outcomes of these conditions differ (3,4). These differences may relate to the degree of intracoronary thrombosis and vasospasm, duration of the acute thrombotic occlusion and ability of the underlying plaque disruption to heal (4). Differences in these outcomes may also relate to differences in study design, in the characteristics of the comparative study samples and to control (or lack thereof) for other demographic or clinical factors that may affect in-hospital and post-discharge outcomes of patients with QWMI and NQWMI. Hospital mortality appears to be greater for patients with QWMI than it is for those with NQWMI (5). On the other hand, the rates of reinfarction, recurrent angina and long-term mortality appear to be higher for hospital survivors of NQWMI (3).
Given the magnitude of coronary heart disease in the U.S., it remains of importance to examine differences in the incidence, in-hospital and long-term outcomes of patients with NQWMI compared with those with QWMI and to determine whether trends in patterns of infarction have changed over time. Using data from the population-based Worcester Heart Attack study, we examined trends in the incidence, hospital and long-term mortality of 5,832 residents of the Worcester, Massachusetts metropolitan area who had been admitted to all greater Worcester hospitals between 1975 and 1997 with confirmed AMI. This 22-year interval spans the periods before and after the establishment of more effective cardiac therapies used in the management of patients with AMI.
The population under study consisted of residents of the Worcester, Massachusetts, metropolitan area hospitalized with a discharge diagnosis of AMI in all acute care general hospitals in the Worcester Standard Metropolitan Statistical Area (SMSA) (1990 census estimate = 437,000) during 1975, 1978, 1981, 1984, 1986, 1988, 1990, 1991, 1993, 1995 and 1997 (6–8). Sixteen teaching and community hospitals were included in this study from the mid 1970s through the late 1980s, while fewer hospitals were included thereafter due to hospital closures, mergers or conversion to chronic care facilities. The medical records of greater Worcester residents with a discharge diagnosis of AMI from these hospitals were individually reviewed and validated by physician and nurse abstractors as previously described (6–8).
The criteria for diagnosis of AMI included at least two of the following: 1) clinical history of prolonged chest pain; 2) serum enzyme elevations in excess of the upper limit of normal as specified by the laboratory at each hospital; and 3) serial electrocardiogram (ECG) tracings during hospitalization showing changes in the ST segment or Q-waves. A diagnosis of QWMI was accepted when Q-waves with a duration of 0.04 s or more and an amplitude equal to or greater than 25% of the R wave in that lead, as well as a typical pattern of evolutionary changes in the ST segment and T waves, appeared in two or more contiguous leads in serial ECGs with or without a typical clinical history or serum enzyme elevations. A diagnosis of NQWMI was accepted when, in the absence of Q-waves, ischemic ST segment and T wave changes persisted for at least 24 h in the presence of enzyme elevations or a typical history. These changes consisted of either ST segment depression ≥1 mm (with or without symmetrical T wave inversion) or symmetrical T wave inversion alone. ST segment elevation AMI was classified as Q-wave or non-Q-wave based on the presence or absence of Q-waves 24 h after hospital admission. These ECGs were interpreted by a trained team of physician and nurse reviewers who received extensive training in the review of ECGs and medical records. Quality control activities on the interpretation of ECGs and abstraction of information from patients’ charts were routinely incorporated over the course of the study. Patients with confirmed AMI who died within minutes of hospital admission, or who died shortly thereafter without the appearance of Q-waves in their ECGs, were classified as NQWMI. Troponin assays were not used to diagnose AMI. Cases of perioperative AMI or those secondary to other noncardiac conditions were not included. The complications of AMI were assessed on the basis of information available from the clinical charts (6–8). The 11 study periods were categorized into five periods of two years each and one single year period to make subsequent data analyses more interpretable.
The hospital records of patients with validated AMI were abstracted for demographic and clinical data including age, gender, medical history, complications during hospitalization, ECG findings and use of diagnostic procedures and therapeutic approaches. Records of previous hospitalizations for coronary heart disease were reviewed where available when the review of the hospital chart indicated that the present hospitalization was not the first for coronary heart disease. This was done to identify initial cases of AMI that constituted the sample of the present report. Patients with ECG changes indicative of prior MI (old Q-waves on ECG) or prior documented MI (confirmed by medical records of prior hospitalizations) were excluded from this study. Approaches used to ascertain survival status after hospital discharge included a review of records for additional hospitalizations and a statewide and national search of death certificates for residents of the Worcester SMSA.
Differences in the distribution of demographic, clinical and treatment characteristics in hospitalized patients according to AMI type (Q-wave vs. non-Q-wave) were examined by use of chi-square tests of statistical significance for discrete variables and ttests for continuous variables. Analysis of variance was used to examine the significance of differences in average age of patients with QWMI and NQWMI over time. The simultaneous effects of several potentially confounding variables that might influence in-hospital survival and two-year mortality after hospital discharge were accounted for by means of a logistic multiple regression technique. The variables controlled for included age, gender, a history of angina, diabetes, hypertension or stroke, AMI location and occurrence of congestive heart failure or cardiogenic shock during admission. These variables were controlled for on the basis of their previously known associations with mortality after AMI and because there were differences in the distribution of these factors between patients with QWMI and NQWMI. A life table approach was used to examine the association of QWMI and NQWMI to long-term survivorship. The Cox proportional hazards regression model was used to examine the association of AMI type to long-term prognosis while controlling for potentially confounding prognostic factors and duration of follow-up.
Characteristics of patients with initial QWMI or NQWMI
Over the 22-year study period (Table 1)and during each study interval (Table 2), patients who developed an initial NQWMI were significantly older and more likely to be women than those who developed an initial QWMI. The mean age and the percentage of women in patients developing either QWMI or NQWMI increased over time (Table 2). Patients with NQWMI were more likely to include those with a history of angina, diabetes, hypertension or stroke, and differences in the prevalence rates of these comorbidities between patients with NQWMI and QWMI persisted over time. Overall, patients with NQWMI were more likely to receive aspirin, beta-adrenergic blocking agents and calcium channel blockers during the acute hospitalization (Table 1). Consistent and increasing trends in the use of beta-blockers and angiotensin-converting enzyme inhibitors and declines over time in the use of calcium channel blockers were observed (Table 2). Patients with QWMI were significantly more likely to be treated with thrombolytic agents overall and during each study period. As expected, we observed marked increases over time in the use of thrombolytic agents in patients with QWMI but not NQWMI (Table 2). Overall, patients with QWMI were more likely to receive a percutaneous coronary intervention but less likely to undergo coronary artery bypass grafting during the index hospitalization (Table 1). However, the absolute magnitude of the difference between our respective comparison groups in the receipt of these procedures was small. The proportion of patients with AMI undergoing a percutaneous coronary intervention increased over time, though this increase was greater in patients with QWMI than it was in patients with NQWMI (Table 2).
Patients with QWMI were more likely to develop congestive heart failure and cardiogenic shock during the acute hospitalization overall as well as during each of the periods under study (Tables 1 and 2).
Temporal trends in the incidence rates of initial QWMI or NQWMI
A multivariable adjusted regression analysis was used to examine trends over time in the likelihood of developing NQWMI while controlling for previously described variables that differed between the respective comparison groups. Relative to the referent period of 1975/78 (odds ratio = 1.0), patients hospitalized in 1981/84, 1986/88, 1990/91, 1993/95 and 1997 were at 1.35 (95% confidence interval [CI]: 1.11, 1.63), 1.75 (95% CI: 1.44, 2.13), 2.09 (95% CI: 1.72, 2.53), 2.91 (95% CI: 2.41, 3.51) and 3.38 (95% CI: 2.72, 4.19) times greater likelihood of developing non-Q-wave AMI, respectively.
The rapid opening of occluded coronary arteries by administration of thrombolytic therapy in patients with ST-segment elevation MI may lead to an increase in the occurrence of NQWMI and a corresponding decrease in the incidence of QWMI by halting or attenuating myocardial damage and preventing the development of Q-waves on the ECG. Therefore, we carried out an additional analysis of trends in the incidence of NQWMI after we excluded patients who received thrombolytic therapy during hospitalization. Despite excluding these patients, the incidence of NQWMI consistently increased over time. For example, the incidence of NQWMI (per 100,000 population) increased from 62 in 1975/78 to 68 in 1986/88, 88 in 1993/95 and 117 in 1997.
Hospital death rates according to type of AMI
Overall, hospital death rates were significantly higher for patients with QWMI (19.5%) as compared with those with NQWMI (12.5%) (p < 0.001). After controlling for previously described demographic and clinical covariates, as well as time period of hospitalization, patients with QWMI remained at significantly increased risk for hospital mortality (adjusted odds of dying = 1.63; 95% CI: 1.35, 1.97).
While the unadjusted hospital mortality rates of QWMI have progressively declined over the 22-year study period, the hospital mortality rates for NQWMI have remained the same (Fig. 2). After adjusting for demographic, medical history and clinical characteristics, there was a consistent and significant decline in the risk of dying for patients with initial QWMI (Table 3). In contrast, the multivariable adjusted odds of dying for patients with initial NQWMI were inconsistent over time and did not change significantly during the most recent study periods (Table 3).
Trends in long-term survival after hospital discharge
Overall, the long-term survival rates for patients with QWMI were 91%, 86%, 75% and 56% at 1, 2, 5 and 10 years after hospital discharge, respectively. Corresponding survival rates were 85%, 80%, 65% and 47% for patients with NQWMI (Fig. 3).
A Cox proportional hazards regression analysis confirmed the lower overall, albeit nonsignificant, long-term mortality for patients with QWMI, as compared with those with NQWMI, after controlling for other prognostic characteristics (adjusted odds ratio = 0.92, 95% CI: 0.84, 1.01).
Since mortality after hospital discharge for AMI is greatest within the first two years and because of the availability of two-year follow-up data for each cohort (with the exception of patients hospitalized in 1997) we examined trends in two-year mortality according to type of AMI for patients discharged between 1975 and 1995. Unadjusted two-year mortality rates after hospital discharge decreased for patients with QWMI (Fig. 4A)but increased for those with NQWMI (Fig. 4B). After simultaneously adjusting for various prognostic factors, the multivariable adjusted two-year mortality rates decreased significantly over time for patients with QWMI. This was particularly evident from the mid-1980s on when thrombolytic therapy and other effective cardiac medications were increasingly utilized in these patients. The two-year mortality rates for patients with initial NQWMI also declined over time, albeit to a nonsignificant extent (Table 4).
The results of this population-based study carried out among residents of a large representative Northeastern metropolitan area demonstrate that the incidence of initial NQWMI has increased markedly over the 22-year period examined. While patients with QWMI experienced significantly higher overall hospital death rates than those with NQWMI, in-hospital and post-discharge mortality decreased in patients with QWMI. In contrast, hospital mortality for patients with NQWMI remained essentially the same over this two-decade long period, while their post-discharge survival improved, particularly during recent study years. The use of medications proven to be effective in the management of patients with AMI increased over time in patients with QWMI and NQWMI. However, the greatest increases in the use of these agents were seen in patients with QWMI, which may have partially contributed to the encouraging trends in prognosis observed in these patients.
A recent editorial (9)suggested that there is no meaningful pathophysiological distinction between QWMI and NQWMI. Phibbs et al. (9)note flaws in previous studies suggesting that the outcomes of patients with QWMI and NQWMI were different. These authors suggested that all studies examining differences in the outcomes of patients with QWMI and NQWMI should include only initial AMIs and control for possible differences in age, infarct location and size. In our study, we controlled for these and other variables. In addition, we have the advantage of being population-based (in contrast with treatment-based) with long-term follow-up over a 22-year study period. Therefore, the results of our study suggest that QWMI is, indeed, a distinctive pathophysiologic entity from NQWMI.
Incidence of NQWMI
Similar to this study, several previous studies (10–12), including an earlier report from the Worcester Heart Attack study (13), suggested an increase in the incidence of NQWMI over varying periods preceding the thrombolytic era. The increased incidence of NQWMI observed in our study may be related to a more frequent use of the ECG for diagnostic purposes or more frequent assaying for cardiac enzymes over the periods under study. However, we do not have data on the frequency of ECG or serial cardiac enzyme use. Our methods of ascertainment and diagnosis of cases of AMI remained the same over time. Thus, while routine use of more sensitive markers of myocardial necrosis has increased the detection of AMI in other studies (14,15), it is not responsible for the increased incidence of NQWMI seen in this study. Changes in the natural history of AMI due to a modification of coronary risk factors, reduction in prehospital delay after the onset of symptoms and improvements in access to and advances in medical care may also have contributed to the trends seen in our study. Indeed, marked changes over time in management practices of patients with AMI have occurred in our study population (16,17).
The use of early reperfusion strategies for ST segment elevation MI has been suggested to play a role in increasing the development of NQWMIs by converting QWMI to NQWMIs (4). Ten to fifteen percent of patients with NQWMI in our study received thrombolytic agents, which is slightly lower than the rates of utilization of these agents seen in other studies (18–24). These usage rates reflect the multi-hospital, community-based perspective of this study. However, in our study, the incidence of NQWMI began increasing before the more widespread use of thrombolytic agents in the late 1980s to early 1990s. In addition, when we carried out a subgroup analysis excluding patients who received thrombolytic therapy, trends toward an increased incidence of NQWMI remained. Thus, conversion of QWMI to NQWMIs by reperfusion therapy can only partially explain the increased incidence of NQWMI seen in our study.
Other factors in addition to the more widespread use of thrombolytic agents may also have contributed to the observed increases in NQWMI. For example, use of aspirin has been shown to limit infarct size and reduce the risk of developing QWMI in patients not receiving thrombolytic therapy (25,26). Marked increases over time in the use of aspirin were observed in our study. In addition, the increasing use of other medications or treatment strategies over the past 22 years may have influenced the incidence rates of NQWMI (8,27).
National (28,29)and population-based (30,31)changes in the commonly accepted risk factors for coronary artery disease, including cigarette smoking and elevated serum cholesterol and blood pressure levels, have demonstrated consistent improvements over the past several decades. These improvements may be associated with an increase in the incidence of NQWMI or “milder” MIs. Unfortunately, changes over time in the major and less well-established risk factors for AMI are not monitored as part of our surveillance activities of the Worcester Heart Attack study. Therefore, we cannot ascertain the effects of risk factor improvements on the incidence of NQWMI in this study.
Given the multitude of potential reasons for the increased incidence of NQWMI demonstrated in our study, we cannot know for certain which of the above factors, if any, is responsible for this increased incidence.
In a review of data from studies carried out between 1962 and 1993 (5), initial NQWMI was associated with a hospital death rate of 6.8%, while QWMI was associated with a hospital case-fatality rate of 15.8%. Our study results confirm the higher overall hospital death rates for patients with initial QWMI over the past two decades.
Despite the adverse overall impact of QWMI on hospital prognosis, marked declines in hospital death rates over time were observed in patients with QWMI, whereas patients with NQWMI experienced little to no change in hospital mortality. As a result of these trends, hospital mortality for patients with QWMI was similar to that for patients with NQWMI in 1997. While the reasons for the encouraging decline in mortality for patients with QWMI are unknown, this reduction likely reflects the increased use of therapies shown to be effective in the management of ST segment elevation MI (8,32), including use of aspirin (33), beta-blockers (34), acute reperfusion strategies (thrombolytic therapy and percutaneous coronary interventions) (35–38)and angiotensin-converting enzyme inhibitors. The 70% reduction in the multivariable adjusted risk of dying during the acute hospitalization for patients with QWMI found in our study is similar to the 60% reduction seen in Registre Gironi del COR (REGICOR) (39), a community-based study of patients with initial QWMI hospitalized in Gerona, Spain, between 1978 and 1993. Despite their significantly lower overall in-hospital mortality, targeted intervention efforts remain needed for patients with NQWMI given their lack of improvement over time in hospital survival between 1975 and 1997. The encouraging results observed with the use of new therapies such as platelet GPIIb-IIIa antagonists (40)for the treatment of acute coronary syndromes may enhance the short- and long-term outlook of patients with NQWMI.
We (8)and others (32)have recently reported encouraging trends in the long-term survival of patients discharged from hospitals after AMI. The results of this study suggest that this benefit is primarily seen in patients with QWMI, although a trend towards improved two-year survival after NQWMI was seen during the most recent study periods. This study is consistent with others (5)that demonstrate a higher long-term mortality for patients with NQWMI compared with those with QWMI.
Patients discharged from the hospital after NQWMI have been shown to have more jeopardized myocardium and a greater concurrent risk of both recurrent ischemia and reinfarction (4,41). Therefore, the improving trends observed in two-year survival rates for patients with QWMI and NQWMI probably reflect improved therapeutic efforts to limit the extent of acute myocardial necrosis, reduce further ischemic burdens and improve left ventricular function.
The use of a survivor cohort (Fig. 4)has implications for interpreting differences in long-term mortality risk. Higher in-hospital mortality for patients with QWMI removes potentially high-risk patients from the post-discharge cohort. Similarly, fewer in-hospital deaths in patients with NQWMI potentially increase the proportion of higher risk patients who are available for long-term follow-up and their associated risk for post-discharge outcomes. These possible caveats need to be kept in mind when interpreting our long-term mortality data.
Study strengths and limitations
This study was carried out among residents of a well-defined metropolitan area whose sociodemographic and economic characteristics reflect those of the U.S. population as a whole, with the exception of race (the vast majority of residents of the Worcester metropolitan area are white). The strengths of this study are its large sample of men and women with validated AMI and its population-based design. Because of the nonrandomized observational nature of this study, we did not control for the use of interventional procedures or medical therapies in examining differences in hospital and post-discharge survival in patients with QWMI or NQWMI. While trends in the incidence and case-fatality rates of Q-wave or NQWMI may be affected by the use of these management approaches, we did not control for the use of these therapies in additional regression analyses. The nonrandomized nature of this study makes analysis of the influence of medical therapies on incidence and case-fatality rates of QWMI and NQWMI difficult to interpret.
Due to our methods of data abstraction, we did not collect data on lifestyle practices or treatment approaches (including revascularization procedures) used by patients after leaving the hospital and, therefore, are unable to examine the impact of these interventions on post-discharge survival. We were also unable to determine whether these treatment approaches were differentially applied according to type of AMI or if competing causes of death differed between patients with QWMI and NQWMI after hospital discharge.
While the trends of improved hospital and post-discharge outcomes for patients with QWMI are encouraging, the lack of improvement over the past 22 years in the hospital survival rates for patients with NQWMI is disappointing. This study underscores the evidence from previous studies that QWMI is associated with greater in-hospital mortality but lower post-hospital mortality than NQWMI, despite the more aggressive approaches to in-hospital management of AMI. Moreover, our data suggest that NQWMIs are likely to continue to account for an increasing proportion of all AMIs. Thus, investigation into the prevention and treatment of NQWMI remains of considerable importance.
☆ Supported by a grant (RO1 HL35434) from the National Heart, Lung and Blood Institute.
- acute myocardial infarction
- confidence interval
- myocardial infarction
- non-Q-wave myocardial infarction
- Q-wave myocardial infarction
- Standard Metropolitan Statistical Area
- Revision received September 20, 2000.
- Revision received January 19, 2001.
- Accepted January 29, 2001.
- American College of Cardiology
- Theroux P,
- Fuster V
- Davies M.J,
- Thomas A.C
- Zaacks S.M,
- Liebson P.R,
- Calvin J.E,
- Parrillo J.E,
- Klein L.W
- Gibson R.S
- Goldberg R.J,
- Yarzebski J,
- Lessard D,
- Gore J.M
- Phibbs B,
- Marcus F,
- Marriott H.J,
- Moss A,
- Spodick D.H
- Edlavitch S.A,
- Crow R,
- Burke G.L,
- Baxter J
- Roberts R,
- Kleiman N.S
- Antman E.M,
- Braunwald E
- Aguirre F.V,
- Younis L.T,
- Chaitman B.R,
- et al.
- Barbagelata A,
- Califf R.M,
- Sgarbossa E.B,
- et al.
- Langer A,
- Goodman S.G,
- Topol E.J,
- et al.
- Matetzky S,
- Barabash G.I,
- Rabinowitz B,
- et al.
- Goodman S.G,
- Langer A,
- Ross A.M,
- et al.
- McGovern P.G,
- Folsom A.R,
- Sprafka J.M,
- et al.
- Havlik R,
- Feinleib M
- Luepker R,
- Higgins M
- Sprafka J.M,
- Burke G.L,
- Folsom A.R,
- Luepker R.V,
- Blackburn H
- Califf R.M,
- Topol E.J,
- Gersh B.J
- Gil M,
- Marrugat J,
- Sala J,
- et al.
- Boersma E,
- Akkerhuis K.M,
- Theroux P,
- Califf R.M,
- Topol E.J,
- Simoons M.L