Author + information
- Received July 5, 1996
- Revision received December 17, 1996
- Accepted January 9, 1997
- Published online April 1, 1997.
- ↵*Dr. Judith K. Mickelson, Cardiology IIIB, Houston Veterans Affairs Medical Center, 2002 Holcombe Boulevard, Houston, Texas 77030.
Objectives. The influence of race and age on thrombolytic therapy, invasive cardiac procedures and outcomes was assessed in a Veterans Affairs teaching hospital. The influence of Q wave evolution on the use of invasive cardiac procedures and outcome was also assessed.
Background. It is not well known how early revascularization procedures for acute myocardial infarction are delivered or influence survival in a Veterans Affairs patient population.
Methods. From October 1993 to October 1995, all patients with myocardial infarction were identified by elevated creatine kinase, MB fraction (CK-MB) and one of the following: chest pain or shortness of breath during the preceding 24 h or electrocardiographic (ECG) abnormalities.
Results. Racial groups were similar in terms of age, time to ECG, peak CK and length of hospital stay. Mortality increased with age (odds ratio [OR] 1.93, 95% confidence interval [CI] 1.33 to 2.81). A trend toward increased mortality occurred for race other than Caucasian. Patients meeting ECG criteria were given thrombolytic agents in 49% of cases, but age, comorbidity count and Hispanic race decreased the probability of thrombolytic use. Cardiac catheterization was performed more often after thrombolytic agents (OR 1.85, 95% CI 0.97 to 3.54), but less often in African-Americans (OR 0.59, 95% CI 0.35 to 1.02), older patients (OR 0.39, 95% CI 0.24 to 0.64) or patients with heart failure (OR 0.30, 95% CI 0.17 to 0.52). Patients evolving non–Q wave infarctions were older and had increased comorbidity counts and trends toward increased mortality. Angioplasty was chosen less for patients ≥65 years old (p = 0.02); angioplasty and coronary artery bypass graft surgery were performed less in patients ≥70 years old (p = 0.02). Patients treated invasively had lower mortality rates than those treated medically (p < 0.02).
Conclusions. The use of thrombolytic agents and invasive treatment plans declined with age, and mortality increased with age. Trends toward increased mortality occurred with non–Q wave infarctions and race other than Caucasian.
(J Am Coll Cardiol 1997;29:915–25)
© 1997 by the American College of Cardiology
Acute myocardial infarction is a major cause of death in the United States. Clinical investigations over the past decade have been numerous, involving thousands of patients worldwide ([1–3]). These trials have provided information that has guided the prompt use of treatment regimens ([4–10]). Early revascularization procedures for acute myocardial infarction, such as administration of thrombolytic therapy or primary angioplasty ([11, 12]), have significantly reduced mortality in eligible patients. However, it is not as well known how these interventions are delivered to varied patient populations in different clinical settings ([13–26]). This study examined the influence of race and age on the use of thrombolytic therapy, invasive cardiac procedures and subsequent outcomes in a Veterans Affairs teaching hospital. The influence of Q wave evolution on the use of invasive cardiac procedures and outcome was also assessed.
1.1 Patient group.
The Houston Veterans Affairs Medical Center is located in a metropolitan area, is affiliated with Baylor College of Medicine and relies on housestaff and subspecialty fellows to provide primary care on the Medical Service. The hospital has 256 beds staffed by the Medical Service. The Emergency Department (emergency room and receiving clinic) is not part of the Medical Service or the medical school programs. During fiscal year 1994 (October 1993 through September 1994), there were 16,603 discharges and 322,207 outpatient visits, including 15,939 emergency room visits and 42,772 receiving clinic (acute care) visits. During fiscal year 1995 (October 1994 through September 1995), there were 16,873 discharges and 331,216 outpatient visits, including 14,415 emergency room visits and 46,788 receiving clinic visits. The racial composition of the patients discharged during the 2-year period included 62% Caucasian, 32% African-American and 6% Hispanic. The information reported was gathered as part of the Quality Assessment and Improvement Program of the Medical Service.
Over a 2-year period (October 1993 through September 1995), charts on all patients admitted to the hospital were reviewed to determine those with the diagnosis of possible myocardial infarction. Patients with myocardial infarction were identified by a higher than normal creatine kinase, MB fraction (CK-MB) (0 to 12 ng/ml) and >5% of the total CK and one of the following: chest pain or shortness of breath during the preceding 24 h or electrocardiographic (ECG) abnormalities.
1.2 Study variables.
For patients identified as having had a myocardial infarction, the ECG was reviewed for ST segment elevation, localization of any change (anterior, inferior, lateral or nonlocalized) and evolution of Q waves. Patient race, age, gender, emergency room arrival time (hour and shift [days: 7 amto 3 pm; evenings: 3 pmto 11 pm; or nights: 11 pmto 7 am]), time that initial ECG was obtained, procedures performed (thrombolytic therapy, cardiac catheterization, coronary angioplasty, coronary artery bypass graft surgery [CABG]), peak CK-MB, length of hospital stay (LOS), discharge medications (aspirin or ticlid, beta-adrenergic blocking agents or amiodarone, calcium channel blockers, long-acting nitrates, angiotensin-converting enzyme [ACE] inhibitors, Coumadin, lipid-lowering drug) and outcome (survival in the hospital and follow-up to July 1996) were recorded. If ST segment elevation was significant enough to warrant evaluation for thrombolytic therapy, the charts were reviewed to determine if treatment was given. If treatment was withheld, the reasons for omission were determined. For those patients treated with thrombolytic therapy, further information was obtained: admission time after emergency room arrival, time from entrance to emergency room to initiation of thrombolytic therapy, time from onset of chest pain to initiation of thrombolytic therapy, type of thrombolytic agent administered and any complications.
Information on comorbid conditions present during the hospital period was obtained from the Patient Treatment File, which is the national data base of information maintained by the Department of Veterans Affairs on all veteran inpatients. This data base includes up to nine secondary diagnoses per hospital admission. Definitions for the following conditions were created from ICD-9-CM codes for medical diagnoses and selected procedures: alcohol and illicit drug use, cancer, cerebrovascular disease (CVD), congestive heart failure, diabetes mellitus, hemiplegia or paraplegia, hypertension, liver disease, peripheral vascular disease (PVD), chronic obstructive pulmonary disease (COPD) and renal disease ([27, 28]).
1.3 Statistical methods.
Data are presented as the mean value ± SEM. When indicated, the median is provided. Comparisons were made by analysis of variance (age, peak CK and CK-MB, time to obtainment of initial ECG, LOS). Simple linear regression analysis was used to compare age with emergency room arrival time, time to obtainment of ECG, time to treatment with a thrombolytic agent, peak CK or LOS. Contingency tables were used to determine if the use of thrombolytic therapy was related to race, age, infarct location, evolution of Q waves, outcome, emergency room arrival time, procedures performed or discharge medications. Follow-up duration was determined by averaging the number of months from the index admission to July 1996 for each patient. A p value <0.05 was considered significant.
Logistic regression was used to examine the relation between patient characteristics and three dependent variables: thrombolytic use in patients with significant ST segment elevation (n = 142), catheterization during the hospital period and in-hospital mortality. Model construction was performed using a hierarchic approach; age and race were included in all models, and individual comorbid conditions were examined if they were univariately associated (p < 0.2) with thrombolytic use, catheterization or in-hospital death, or were of strong clinical interest. In addition, a simple count of the number of comorbid conditions was also examined in the models. The comorbidity count was handled as a continuous variable; other researchers have found that it has predictive power for patient outcomes (). However, because the presence of congestive heart failure alone was strongly associated with the use of catheterization, it was excluded from the comorbidity count for the logistic models only and entered separately in each of these models. Age and comorbidity count were examined both as continuous variables and as categoric ones with dummy variable coding; when the coefficients in the latter model confirmed a linear relation between age or count of comorbid conditions and the dependent variable of interest, they were retained as continuous variables (). In the model predicting thrombolytic use and catheterization, increasing age had little effect until it exceeded 70 years; thus, it was handled in these analyses as a dichotomous variable with that cutoff point. All results were reported as odds ratios (OR) with 95% confidence intervals (CI). All logistic regression analyses were performed using the SAS package.
2.1 Incidence of myocardial infarction.
Over the 2-year study period, 1,703 patients were admitted to the Coronary Care Unit, 685 were evaluated for possible myocardial infarction and 355 developed myocardial infarctions. Two women were not included in further analysis. Myocardial infarction occurred in the hospital in 82 patients. There were no differences among the shifts when patients presented to the emergency room with respect to the number of patients, time to obtainment of ECG or age. The time of emergency room presentation was used to visualize the number of patients arriving during each hour. There was a trend toward a circadian rhythm with a primary peak incidence in the midmorning and a secondary peak in the evening (Fig. 1).
2.2 Racial groups.
There was no difference among the racial groups in terms of age, time to obtainment of initial ECG, peak CK or CK-MB, LOS or infarct location (Table 1). The anterior and inferior infarcts had higher peak CK and CK-MB levels than the lateral and nonlocalized infarcts (p = 0.008). There were no differences among infarct locations in terms of time to obtainment of initial ECG, LOS, subsequent cardiac catheterization, percutaneous transluminal coronary angioplasty (PTCA) or CABG. For the entire study group, there was no difference among the racial groups with respect to being treated with thrombolytic therapy. Aspirin was more frequently prescribed for Caucasians, and this persisted even if patients taking Coumadin were excluded from the analysis. There was no difference among the racial groups in the use of beta-blockers (52.3%), calcium channel blockers (40.8%), long-acting nitrates (63.4%), ACE inhibitors (48.4%), Coumadin (11.8%) or lipid-lowering drugs (20.2%). Patients who did not survive to discharge were older and had more comorbid diseases than those who survived (69.6 ± 1.1 vs. 62.4 ± 0.6 years, p = 0.0001) (Table 2). There was no difference in total comorbidity count, but each racial group was unique in terms of the prevalence of comorbid conditions. There were trends toward increased mortality for race other than Caucasian, which may have been related to the particular comorbidity for Hispanics (diabetes and CVD) and African-Americans (hypertension and heart failure). At a median of 22 months of follow-up, mortality had doubled overall, with less differences among the racial groups.
2.3 Age groups.
If older patients were considered to be those ≥65 years old, they differed from their younger cohort in several ways: mortality was higher; thrombolytics were used less; lateral or nonlocalized non–Q wave infarctions occurred more often; catheterization and angioplasty were performed less often; and the comorbidity count increased (Table 3, Fig. 2). Patients 75 years and older were not hospitalized longer, nor did they have higher peak CK or CK-MB levels than younger patients. They did, however, wait longer for the initial ECG and they were less likely to receive beta-blockers, PTCA or CABG than younger patients.
2.4 Use of thrombolytic therapy.
There were 142 patients with ST segment elevation, meeting the ECG criteria for thrombolytic therapy, and 70 were treated. Reasons cited for choosing not to administer thrombolytic therapy varied slightly among racial groups (Table 4). African-Americans had more hypertensive and cerebrovascular exclusions, whereas Hispanics more often presented in cardiogenic shock or were treated with primary PTCA. Hispanics were the smallest patient subgroup, but they were most likely to present with ST elevation and evolve Q waves (Table 1). Hispanics and older patients meeting the ECG criteria for thrombolytic therapy were less often treated (Table 2).
Patients with ST segment elevation who were treated with a thrombolytic agent were younger and more likely to undergo cardiac catheterization and to evolve Q waves than those with ST segment elevation who were not treated (Table 4). Anterior ST segment elevation and infarction were more common among African-Americans. There were no differences among the racial subgroups of patients with ST segment elevation regarding time to obtainment of initial ECG, peak CK and CK-MB, LOS, subsequent PTCA or CABG. There were trends toward decreased mortality for the patients with ST segment elevation who were treated with thrombolytic agents compared with those who were not (Caucasian: 2 [4.3%] vs. 4 [10.8%]; African-American: 1 [6.3%] vs. 3 [15.8%]; Hispanic: 2 [29%] vs. 5 [30%]), except in the small group of Hispanics studied. The patients with ST segment elevation who died before discharge were older than those who survived (69.6 ± 0.8 years [n = 17] and 60.9 ± 1.0 years [n = 125], respectively [p = 0.03]). Patients treated with thrombolytic agents who died did not experience delays in initiation of therapy compared with survivors (presentation to treatment: 93 ± 15 min [median 85] [n = 5] vs. 120 ± 10 min [median 98] [n = 65]; chest pain to treatment: 187 ± 27 min [median 191] [n = 5] vs 271 ± 21 min [median 212] [n = 65], respectively). Mortality after thrombolytic therapy tended to be higher for non–Q wave than for Q wave infarction (in-hospital: 4.3% vs. 12.5%, p = NS; 22 months follow-up: 8.7% vs. 25%, p = 0.06, respectively), and these patients who died were older.
There was no correlation between patient age and duration of chest pain before presentation or time from onset of chest pain or emergency room arrival to initiation of thrombolytic therapy. The time from the onset of chest pain or from presentation to initiation of treatment was not different among racial groups, infarct locations or with Q wave evolution. There was a trend toward a progressively longer time from presentation to treatment with thrombolytic therapy from the day shift to the night shift (Table 5). There was a significantly longer time from the onset of chest pain to initiation of treatment between the day shift and either the evening or night shift. As part of this time delay, there was a longer time from onset of chest pain to presentation between the day shift and either the evening or night shift. For patients treated with thrombolytic agents, there was no difference between shifts with respect to the number of patients, race, outcome or age.
The choice of thrombolytic agent included front-loaded recombinant tissue-type plasminogen activator (rt-PA) (n = 58), streptokinase (n = 8) and reteplase (n = 4). Complications occurred in five patients (one Caucasian, two African-Americans and two Hispanic) 57 to 72 years old (mean [±SEM] 65.6 ± 2.7). No patient required transfusion. With rt-PA one patient each developed epistaxis, gingival bleeding, hand hematoma and embolic stroke. There was one fatal intracranial bleed with streptokinase. The two patients with cerebrovascular events underwent emergent computed tomography and had a remote history of stroke (>2 years).
2.5 Thrombolytic therapy versus conventional therapy.
When the entire study group was evaluated, the majority of patients who developed myocardial infarction were not candidates for thrombolytic therapy (n = 283). Patients treated with thrombolytic therapy (n = 70) were younger, the initial ECG was obtained faster and peak CK and CK-MB were higher (p < 0.001). There were more anterior and inferior Q wave infarcts in patients treated with thrombolytic agents (p < 0.001). Cardiac catheterization (p < 0.001) and PTCA (p < 0.05) were performed more often in patients receiving thrombolytic therapy. The comorbidity count and age >70 years both tended to decrease the probability of thrombolytic use (Table 2). Patients treated with thrombolytic agents were more often discharged on beta-blockers (p < 0.001). In-hospital mortality, when selected for thrombolytic therapy, tended to be lower than for conventional therapy (7.1% vs. 13.8%, p = 0.10), and this survival benefit persisted long term (14.3% vs. 28.6%, p < 0.01).
2.6 Non–Q wave versus Q wave infarction.
The patients with non–Q wave infarctions were older, had lower peak CK and CK-MB levels and were less likely to undergo PTCA than patients with Q wave infarctions, but the use of cardiac catheterization and CABG was not different (Table 6). The incidence of non–Q wave infarction was more common among Caucasians and African-Americans than Hispanics (Table 1). Evolution of a non–Q wave infarction was more likely in patients without ST segment elevation or thrombolytic therapy and included more lateral and nonlocalized infarcts. Patients with non–Q wave infarctions were more often discharged on calcium channel blockers and ACE inhibitors, whereas patients with Q wave infarctions were discharged on beta-blockers. The simple comorbidity count, prevalence of problematic comorbid conditions and mortality tended to be higher in patients with non–Q wave infarctions.
2.7 Invasive versus noninvasive treatment plans.
Cardiac catheterization was performed in 210 patients. These patients were younger and were more often treated with thrombolytic agents (p < 0.001). However, LOS was longer (p < 0.05). African-Americans were less likely to undergo cardiac catheterization, as were older patients (Table 2). Heart failure decreased the probability of cardiac catheterization. Patients undergoing cardiac catheterization were more often discharged on aspirin, beta-blockers and calcium channel blockers (p < 0.001), whereas patients who did not have catheterization were more often discharged on ACE inhibitors (p < 0.05). In-hospital mortality, when selected for cardiac catheterization, was lower than for noninvasive therapy (4.8% vs. 24.1%, p < 0.001), and this survival benefit persisted long term (13.8% vs. 43.2%, p < 0.001).
The subsequent treatment plans—PTCA versus CABG versus medical therapy—differed in several ways. Patients undergoing PTCA were younger, more often had ST segment elevation and developed Q waves (p < 0.01). Coronary angioplasty was performed less often in patients ≥65 years old than in the younger cohort (Table 3). The three deaths after PTCA occurred more than 4 days after the procedure in older patients with previous CABG and severe left ventricular dysfunction. Bypass graft surgery was not frequently pursued in the immediate postinfarction period; however, all these patients survived the initial hospital period despite having the highest comorbidity count (Table 7). Patients with Q wave infarctions did not undergo surgical treatment for at least 7 days after admission, accounting for the longer hospital period with this treatment plan. Both PTCA and CABG were performed less often in patients ≥75 years old. These patients selected for invasive treatment plans had lower mortality rates than those treated medically.
3.1 Incidence of myocardial infarction.
The acute coronary syndromes (myocardial infarction and unstable angina) are a major health concern in the United States with regards to mortality, morbidity and health care delivery. Use of thrombolytic therapy and invasive treatment for acute myocardial infarction in a Veterans Affairs Medical Center affiliated with a medical school were scrutinized in this study. The potential study group—patients admitted with the diagnosis of possible myocardial infarction (2.0%)—and the final study group—patients with myocardial infarction (1.0%)—represent a small percentage of the patients discharged from the hospital during the 2-year period. Patients who were candidates for thrombolytic agents and those who were subsequently treated represent a fraction of the population. Patients with myocardial infarction had a high in-hospital mortality rate. Mortality in the hospital and at a median of 22 months of follow-up was better in patients selected for thrombolytic therapy or invasive cardiac procedures (catheterization, PTCA or CABG), or both.
3.2 Racial groups.
Many of the large studies, when racial composition is reported, are >90% Caucasian ([15, 16, 31–33]). Our study group is unique when compared with both these large studies and smaller studies in public hospitals (). Despite controlling for age, comorbidity count and heart failure, Hispanic patients meeting the ECG criteria for thrombolytic therapy were less often treated than Caucasians or African-Americans. However, the two primary angioplasties performed were in Hispanic men. Overall, our Hispanic population included older men with large infarctions and an increased incidence of cardiogenic shock, which accounts for the poor survival in this particular group. Our study evaluated the use of invasive procedures in a very specific patient group, those with acute myocardial infarction during the index hospital period. This group of patients should be more homogeneous with respect to treatment among racial groups than other ischemic syndromes, yet there were significantly fewer African-Americans who underwent cardiac catheterization. This persisted despite controlling for trends toward increased hypertension and heart failure in African-Americans. There appears to be a well-documented trend toward less invasive strategies in African-Americans ([21–26, 31, 35]), which in our patients is associated with a trend toward increased mortality. However, one recent retrospective analysis within the Department of Veterans Affairs found that although African-Americans did not undergo catheterization or revascularization procedures as often after a myocardial infarction, their survival was similar to Caucasians (). Thus, invasive treatment plans may not provide a better outcome or improved survival for all patients. However, there is little clinical trial data for racial groups other than Caucasians to guide management ([31–33, 36–40]). There was also a difference in medical therapy among the races; aspirin was prescribed more often for discharged Caucasians than for the other racial groups. Unfortunately, specific data on conditions such as gastrointestinal bleeding, structural intracranial abnormalities and severe hypertension that may have prevented or hindered the use of aspirin were not obtained. Our study group was too small to comment on differences in the use of PTCA or CABG among the racial groups.
3.3 Age groups.
As in our group, other investigators () have found that mortality in myocardial infarction increases with age, whereas the probability of receiving thrombolytic agents or an invasive treatment plan decreases. However, when thrombolytic therapy or an invasive treatment plan was chosen, survival improved, regardless of age. In these older patients, when thrombolytic agents are contraindicated, a different approach will be required to improve survival (i.e., perhaps earlier cardiac catheterization and revascularization procedures).
3.4 Use of thrombolytic therapy.
In this particular study group, 19.8% of the patients with symptoms within the past 24 h who developed a myocardial infarction received thrombolytic therapy. This is lower than the 35.1% reported from the National Registry of Myocardial Infarction () but not different from the Myocardial Infarction Triage and Intervention (MITI) project () (Table 8was constructed to facilitate comparison of our data with other studies). Patients with diagnostic ECG changes (localized ST segment elevation) were treated in 49% of the cases, similar to the 52% treated in Global Utilization of Streptokinase and TPA for Occluded Arteries (GUSTO). If a more standard time window were used (chest pain <12 h), we treated 58% of patients with diagnostic ECG changes. By liberal guidelines, including chest pain duration up to 24 h and regardless of age, 65.4% of eligible patients without significant contraindications were treated. The remainder of eligible patients were not treated because the ECG changes were not appreciated or the housestaff and cardiology fellow were unable to obtain, or the patient was unable to provide, a good history. Most often the patient would describe vague, intermittent chest pain over the preceding several days, despite having significant ST segment elevation at presentation. During the study period, there was one fatal bleeding incident—an intracranial bleed after administration of streptokinase (1.4%). It may be difficult in a teaching hospital with this particular patient group to increase the percentage of patients safely treated with thrombolytic agents.
The majority of patients were treated with front-loaded rt-PA (82.9%). Mortality in the treatment group is within the range of the various study groups in the GUSTO trial, but it is probably higher than the 5.9% reported by the National Registry of Myocardial Infarction or the 4.9% reported by the Thrombolysis in Myocardial Infarction (TIMI) phase II trial ([1, 13, 36]) (Table 8). These mortality differences are most likely related to our small sample size, the 24-h time window for treatment and the increased number of older non-Caucasian patients. In a racial subgroup analysis of the TIMI II data, the African-American and Hispanic men were more than a decade younger than our patients, and inclusion criteria required presentation within 4 h of chest pain onset (). We did not find a difference in time to initiation of therapy and the outcome among patients who received thrombolytic therapy, but our time to initiation of therapy after the onset of chest pain was probably longer than that in reported studies. Any delay in treatment is to be avoided, because it is well recognized that the earlier treatment can be initiated, the lower the mortality ([42–45]).
Recent studies to define delays in initiating treatment included time spent obtaining an ECG by personnel not present in the emergency department, bedside cardiology consultation and preparation of the infusion in an off-site pharmacy ([46–48]). In our study, electrocardiography was performed by a technician who had to be paged to the emergency department, and all patients were evaluated by the cardiology fellow before treatment. Patients presenting during the evening and night shifts took more time getting to the hospital, and there was a progressively longer time from presentation to treatment. Nighttime delays in the administration of thrombolytic therapy were also reported in the National Registry of Myocardial Infarction (). Although the number of veterans presenting with acute myocardial infarction was similar during each shift of the day, there was a trend toward a circadian rhythm at presentation ([49, 50]).
3.5 Non–Q wave versus Q wave infarction.
Evolution of non–Q wave infarction after thrombolytic therapy occurred in over one-third of our patients, slightly higher than has been reported in other studies ([33, 51]). The mortality rate for non–Q wave infarction after thrombolytic therapy was low and tended to be lower than that for Q wave infarctions at 1 () or 2 years () in these studies. In our small study group, mortality after thrombolytic therapy tended to be higher for non–Q wave infarctions, but these patients were also older. The majority of non–Q wave infarctions in our study occurred in patients who did not have ST segment elevation or receive thrombolytic therapy. Recent information suggesting elderly patients with non–Q wave infarctions have a lower in-hospital mortality rate than those with Q wave infarctions was not confirmed in our study group (). Also, in our study group, non–Q wave infarctions were predominant and these patients were older and had more comorbid diseases than patients with Q wave infarctions. Mortality, both in the hospital and at 22 months of follow-up, in patients with non–Q wave infarctions was high and remained nearly twice that of patients with Q wave infarctions.
3.6 Invasive treatment plans.
Many of the patient characteristics reported in the National Registry of Myocardial Infarction are comparable to our study patients (Table 8). Patients treated with thrombolytic agents were younger, had localizing ECG changes and showed improved survival. As with our patients, invasive procedures were performed more often in patients receiving thrombolytic therapy (). These rates for procedures after thrombolytic therapy are similar to the GUSTO-1 substudy () or the Fee-for-Service hospitals in the MITI project (). Our use of angiography and revascularization procedures early after myocardial infarction was similar to that in U.S. hospitals in several reports ([16, 17]). In the population ≥65 years old with acute myocardial infarction, our use of angiography and revascularization procedures was similar to that in the Medicare population reported for Texas (). Coronary angioplasty was performed less often in patients 65 years or older. Bypass graft surgery was performed less often in patients 70 years or older. Patients over 65 years old had higher comorbidity counts, which may have contributed to a selection bias away from invasive treatment plans. Patients selected for cardiac catheterization, which requires a cardiologist in the patient’s care, were discharged on a more aggressive antianginal regimen than patients who did not undergo cardiac catheterization. However, the prevalence of heart failure was higher in patients who did not have catheterization, a reasonable explanation for using ACE inhibitors rather than calcium channel blockers and possibly beta-blockers. Both groups used more ([13, 16, 17, 53]) or the same amount () of aspirin and beta-blocker reported previously. Within the Veterans Affairs Medical Centers, there are no direct financial incentives or insurance issues affecting the use of procedures or medications.
Hospital stay was 3 days longer in patients undergoing cardiac catheterization. This was in part due to delays in deciding the patients’ need to have cardiac catheterization performed, based on either information obtained from the echocardiogram, stress test or clinical scenario. To train housestaff and fellows in both the thought processes and techniques, a more liberal use of time and testing may be undertaken in a serial rather than a parallel fashion. Unless the procedure is urgent or emergent it will be carried out during routine hours of a 5-day week. The decision to have personnel available in the cardiac noninvasive laboratory, nuclear medicine laboratory or cardiac catheterization laboratory at all times to perform studies reflects local administrative and economic strategies ([18, 19, 32, 54]).
3.7 Study limitations.
This was not a randomized trial of invasive strategies. We cannot conclude that improved survival among patients treated with thrombolytic agents or revascularization, or both, was due to the treatment itself. We do not know whether patients were chosen for less invasive medical therapy based on 1) sound medical judgment, which determined a poor outcome to be inevitable; 2) the patients’ preferences for the noninvasive regimen; or 3) the physicians’ preferences regarding the provision of invasive therapies. Those patients who did not survive to hospital discharge were older and had more heart failure than those treated with thrombolytic agents or those undergoing cardiac catheterization. There were no data collected on in-hospital medications, which may have had an impact on survival.
Given the small number of patients treated, the data available might not be sufficient to conclude that thrombolytic treatment unequivocally improves outcomes for each racial or age group. In particular, Hispanics tended to have more problematic comorbid diseases (i.e., CVD and diabetes). Among Hispanics meeting ECG criteria for thrombolytic therapy, there was no survival benefit to treatment. Bypass grafting was chosen for a significant percentage of Hispanics, which reflects recent practices in the choice of revascularization procedure for diabetics. In contrast, those African-Americans with ST segment elevation who were given thrombolytic agents had a low mortality. The data base on comorbid diseases may be incomplete and thus not accurately display a potential explanation for increased mortality within a racial group. The few patients over 75 years old who were treated with thrombolytic agents did very well, but studies with large numbers of elderly patients have determined that mortality in this age group is still very high ().
There appears to be a relatively large percentage of older patients with non–Q wave infarctions who had several comorbid diseases and an associated high mortality. Chart review yielded vague or no data regarding duration of symptoms before presentation, unless thrombolytic agents were administered. Some of these older patients with non–Q wave infarctions may have presented to the hospital later with serious complications (i.e., heart failure), which hindered any treatment plan that might have improved survival.
We limited our analysis to revascularization during the index hospital period for acute myocardial infarction. Some patients undoubtedly underwent cardiac catheterization and revascularization during a subsequent admission. However, it is unlikely that the lower rate of catheterizations among African-Americans would be completely offset by an excess of these procedures in the months after discharge. Moreover, some patients, particularly those with Medicare, may have obtained services outside of the Veterans Affairs Medical Center, making complete ascertainment of invasive procedures among our patients difficult. Similarly, our outcome information may be incomplete if patients died outside of a Veterans Affairs Medical Center and we were not notified of the event.
We have shown that a busy teaching hospital in a large metropolitan center can safely provide invasive cardiac services at rates comparable to institutions that participated in major investigations and randomized trials. Two-year follow-up in our patient group shows a very high mortality rate despite medical regimens that are as good as those published in large clinical trials. Our study group included elderly and more ethnically diverse patients than can be found in most major published reports. Efforts to improve our care of patients with acute myocardial infarction could include 1) aggressively educating emergency staff, housestaff and cardiology fellows in ECG interpretation and chest pain history taking so that more eligible patients will be offered thrombolytic therapy or primary angioplasty; 2) improving the time to thrombolytic treatment by developing a “thrombolytic team”; 3) vigilance regarding treatment of changing risk factors and disease progression; and 4) continued patient education of the population at risk regarding the need to present to the hospital as quickly as possible after the onset of symptoms. This last task may be difficult to achieve because over one-fourth of the study group resides in Texas counties other than that of the Houston Veterans Affairs Medical Center, making transportation time an issue. Nevertheless, hospitals determined to improve the cardiac care they provide would do well to begin with an examination of their patients with acute myocardial infarction, as we have done. It is possible that other hospitals have patients and providers with unique characteristics that may or may not be comparable to the major published studies of acute myocardial infarction management available to date.
☆ This work was supported in part by grants from the American Heart Association Established Investigatorship (Dr. Mickelson) and American Heart Association Texas Affiliate Grant-in-Aid (Dr. Mickelson) and a Research Associate Career Development Award from the Health Services Research and Development Service, Department of Veterans Affairs (Dr. Geraci).
- angiotensin-converting enzyme
- coronary artery bypass graft surgery
- creatine kinase
- chronic obstructive pulmonary disease
- cerebrovascular disease
- length of hospital stay
- percutaneous transluminal coronary angioplasty
- peripheral vascular disease
- recombinant tissue-type plasminogen activator
- Received July 5, 1996.
- Revision received December 17, 1996.
- Accepted January 9, 1997.
- The American College of Cardiology
- ISIS-3 (Third International Study of Infarct Survival) Collaborative Group
- ACC/AHA Task Force
- (1995) Interventions in the treatment of acute myocardial infarction. J Am Coll Cardiol 25(Suppl 7):1–51.
- for the participants in the National Registry of Myocardial Infarction,
- Maynard C,
- Weaver WE,
- Lambrew C,
- Bowlby LJ,
- Rogers WJ,
- Rubinson RM
- for the participants in the National Registry of Myocardial Infarction,
- Rogers WJ,
- Bowlby LJ,
- Chandra NC,
- et al.
- Carroll RJ,
- Horn SD,
- Soderfeldt B,
- James BC,
- Malmberg L
- Borowsky SJ,
- Kravitz RL,
- Laouri M,
- et al.
- Whittle J,
- Conigliaro J,
- Good CB,
- Lofgren RP
- Hosmer JA,
- Lemeshow S
- for the MITI Investigators,
- Every NR,
- Fihn SD,
- Maynard C,
- Martin JS,
- Weaver WD
- for the TIMI II Investigators,
- Aguirre FV,
- Younis LT,
- Chaitman BR,
- et al.
- Maynard C,
- Fisher LD,
- Passamani ER,
- Pullum T
- TIMI Investigators,
- Taylor HA,
- Chaitman BR,
- Rogers WJ,
- et al.
- Pell ACH,
- Miller HC,
- Robertson CE,
- Fox KAA
- Matetzky S,
- Barabash GI,
- Rabinowitz B,
- et al.
- Chung MK,
- Bosner MS,
- McKenzie JP,
- Shen J,
- Rich MW
- Kennedy HL,
- Rosenson RS