Reports from large studies using administrative data sets and event registries have characterized recent temporal trends and treatment patterns for acute myocardial infarction. However, few were population based, and fewer examined differences in patterns of treatment for patients presenting with ST-segment elevation myocardial infarction (STEMI) and non–ST-segment elevation myocardial infarction (NSTEMI). The aim of this study was to examine 22-year trends in the use of 10 medical therapies and procedures by STEMI and NSTEMI classification in 30,986 definite or probable myocardial infarctions in the Atherosclerosis Risk in Communities (ARIC) Community Surveillance Study from 1987 to 2008. Weighted multivariate Poisson regression, controlling for gender, race and center classification, age, and Predicting Risk of Death in Cardiac Disease Tool score, was used to estimate average annual percentage changes in medical therapy use. From 1987 to 2008, 6,106 hospitalized events (19.7%) were classified as STEMIs and 20,302 (65.5%) as NSTEMIs. Among patients with STEMIs, increases were noted in the use of angiotensin-converting enzyme inhibitors (6.4%, 95% confidence interval [CI] 5.7 to 7.2), antiplatelet agents other than aspirin (5.0%, 95% CI 4.0% to 6.0%), lipid-lowering medications (4.5%, 95% CI 3.1% to 5.8%), β blockers (2.7%, 95% CI 2.4% to 3.0%), aspirin (1.2%, 95% CI 1.0% to 1.3%), and heparin (0.8%, 95% CI 0.4% to 1.3%). Among patients with NSTEMIs, the use of angiotensin-converting enzyme inhibitors (5.5%, 95% CI 5.0% to 6.1%), antiplatelet agents other than aspirin (3.7%, 95% CI 2.7% to 4.7%), lipid-lowering medications (3.0%, 95% CI% 1.9 to 4.1%), β blockers (4.2%, 95% CI 3.9% to 4.4%), aspirin (1.9%, 95% CI 1.6% to 2.1%), and heparin (1.7%, 95% CI 1.3% to 2.1%) increased. Among patients with STEMIs, decreases in the use of thrombolytic agents (−7.2%, 95% CI −7.9% to −6.6%) and coronary artery bypass grafting (−2.4%, 95% CI −3.6% to −1.2%) were observed. Similar increases in percutaneous coronary intervention and decreases in the use of thrombolytic agents and coronary artery bypass grafting were noted among all patients. In conclusion, trends of increasing use of evidence-based therapies were found for patients with STEMIs and those with NSTEMIs over the past 22 years.
The availability of medical therapies for in-hospital management of acute myocardial infarction (AMI) is increasing annually, and the use of these therapies has substantially contributed to decreasing AMI death rates over the past 3 decades. Reports from large observational studies have characterized recent temporal trends and treatment patterns for AMI. Conclusions from existing reports have been limited, however, by selection, short follow-up periods, racially and/or geographically homogenous populations, and unvalidated clinical data. Furthermore, few studies have examined differences in temporal trends in the treatment of patients presenting with ST-segment elevation myocardial infarction (STEMI) compared to non–ST-segment elevation myocardial infarction (NSTEMI), especially since this redefinition of acute myocardial infarction (MI) resulted in a divergence in treatment recommendations by MI subclass beginning in 2000. In this report, we characterize temporal trends in the in-hospital treatment of patients with STEMIs and those with NSTEMIs over a 22-year period in the Atherosclerosis Risk in Communities (ARIC) Study.
Methods
The design of the community surveillance component of ARIC has been described. Briefly, it is a continuous, retrospective surveillance study of hospitalized coronary heart disease events with mortality follow-up designed to estimate trends in coronary heart disease incidence and mortality using standardized criteria and methods in 4 United States communities: Forsyth County, North Carolina; Jackson, Mississippi; 8 suburbs of Minneapolis, Minnesota; and Washington County, Maryland. Eligible events included inpatient and out-of-hospital deaths due to coronary heart disease and hospitalized nonfatal MI in residents of these communities aged 35 to 74 years. Trained abstractors investigate hospitalizations randomly sampled from annual discharge lists obtained from each hospital serving the 4 ARIC communities. Events were sampled on age, gender, community of residence, and International Classification of Diseases, Ninth Revision, discharge codes, including 402, 410 to 414, 427, 428, and 518.4. Collected data items included presenting symptoms; timing of symptom onset; history of MI, angina, and other cardiovascular conditions; in-hospital medications, diagnostics, and medical procedures; laboratory values for a number of relevant cardiac biomarkers; and up to 3 sets of 12-lead electrocardiographic readings. Regular and ongoing interabstractor agreement is assessed by evaluating concordance between data elements from a sample of cases abstracted independently by 2 abstractors. Internal quality control procedures at the electrocardiographic reading center were used to ensure reproducibility.
A computerized algorithm using electrocardiographic readings, history of chest pain, and cardiac biomarker levels (total creatinine phosphokinase, creatinine phosphokinase-MB, lactate dehydrogenase, troponin I, and troponin T) was used to assign an MI diagnosis to sampled hospitalized events. This analysis was restricted to events with definite or probable MI diagnoses. Any event with abnormal or equivocal biomarker levels was further classified as STEMI or NSTEMI using pain presentation and Minnesota-coded electrocardiographic data from the first, third, or last electrocardiogram obtained during hospitalization. Multiple hospitalizations occurring within 28 days were combined and treated as 1 event. Any event requiring review (e.g., events for which the computer-derived classification of definite MI disagreed with the International Classification of Diseases, Ninth Revision, Clinical Modification codes for discharge diagnosis) was independently classified by two trained reviewers. Any disagreements in diagnoses were then adjudicated by a third reviewer.
Medications and procedures were obtained from hospital pharmacy records and medical record review during the abstraction process. Our analysis included data on 7 medication classes (aspirin, β blockers, calcium channel blockers, angiotensin-converting enzyme [ACE] inhibitors, lipid-lowering medications, antiplatelet agents other than aspirin, and heparin) and 4 revascularization procedures (coronary artery bypass grafting [CABG], thrombolytic therapy [intracoronary or intravenous streptokinase, urokinase, anistreplase, anisoylated plasminogen streptokinase activator complex, or tissue plasminogen activator], and coronary angioplasty [PCI] with or without the implantation of a stent). Each medication or procedure was classified as any receipt during hospitalization or at discharge (yes or no). Because abstraction of several therapies of interest began after 1987, trends for the following therapies were estimated beginning with the first study year for which complete treatment information was available for all sampled events: heparin (beginning in 1992), ACE inhibitors (1992), antiplatelet agents other than aspirin (1997), lipid-lowering medications (1999), and stent implantation (1999).
Patient demographics were obtained from medical record reviews. Demographics of interest included gender (male or female), race (black or white or other), and age. Clinical co-morbidities including previous MI, hypertension, diabetes mellitus, and stroke were collected. To adjust for disease severity and clinical co-morbidities, we used a modified Predicting Risk of Death in Cardiac Disease Tool (PREDICT) score. The score is a validated metric that predicts mortality in patients with acute coronary syndromes from clinical presentation data, including cardiogenic shock, history of MI or cardiac procedures, age, severity of electrocardiographic changes, congestive heart failure, and Charlson co-morbidity index. Data on renal function were not collected and therefore were omitted from our PREDICT score calculation.
We excluded patients whose race was not classified as black or white (n = 658) and, because of insufficient sample sizes, black patients who were sampled in Minnesota and Washington County, Maryland (n = 493). After these exclusions, the final sample size for analysis was 30,986 definite or probable MI events.
All estimates presented are weighted to account for the ARIC surveillance sampling scheme. We examined changes in study population characteristics over the study period using chi-square tests for independence with robust variance estimation to account for the complex sampling scheme. The proportion of patients receiving each medication and procedure was calculated for all study years using weighted Poisson regression, with estimates age standardized to the 2000 United States census age distribution. We used multivariate log-linear regression to estimate average annual percentage increases or decreases for each medical therapy overall and among patients with STEMIs and those with NSTEMIs. In the figures, we present medication and procedure use for each study year; however, for ease of reporting and to promote stability in confidence interval estimates, events were grouped into intervals of 5, 6, or 7 years for table presentation. To account for the complex sampling scheme, all analyses were conducted using SAS-callable SUDAAN release 9.2 (Research Triangle Institute, Research Triangle Park, North Carolina).
Results
Table 1 lists selected study population characteristics over time in 5-year intervals. From 1987 to 2008, 30,986 definite or probable MI events were sampled in the 4 study communities. Of these, 6,106 (19.7%) were classified as STEMIs and 20,302 (65.5%) as NSTEMIs. The proportion of events classified as neither STEMIs nor NSTEMIs (14.8%) remained stable over the study period, as did the gender and age distributions. The prevalence of hypertension, diabetes, and stroke all increased throughout the study period. The mean length of stay in days decreased.
| Variable | Overall | 1987–1991 † | 1992–1996 | 1997–2001 | 2002–2008 |
|---|---|---|---|---|---|
| (n = 30,986) ∗ | (n = 7,524) | (n = 7,730) | (n = 7,380) | (n = 8,352) | |
| Age (yrs), mean ± SD | 60.4 ± 12.0 | 61.1 ± 10.9 | 60.6 ± 11.7 | 60.5 ± 11.9 | 59.6 ± 13.7 |
| Men | 20,360 (65.7%) | 4,940 (65.7%) | 5,207 (67.4%) | 4,813 (65.2%) | 5,400 (64.7%) |
| Race-by-center classification | |||||
| Forsyth County black | 3,273 (10.6%) | 616 (8.2%) | 777 (10.0%) | 832 (11.3%) | 1,047 (12.5%) |
| Forsyth County white | 8,889 (28.7%) | 2,103 (28.0%) | 2,311 (29.9%) | 2,130 (28.9%) | 2,345 (28.1%) |
| Jackson black | 3,805 (12.3%) | 621 (8.3%) | 752 (9.7%) | 1,005 (13.6%) | 1,427 (17.1%) |
| Jackson white | 3,276 (10.6%) | 1,145 (15.2%) | 929 (12.0%) | 681 (9.2%) | 522 (6.3%) |
| Minnesota whites | 6,320 (20.4%) | 1,584 (21.1%) | 1,566 (20.3%) | 1,407 (19.1%) | 1,762 (21.1%) |
| Washington County white | 5,422 (17.5%) | 1,454 (19.3%) | 1,396 (18.1%) | 1,324 (17.9%) | 1,248 (15.0%) |
| Co-morbidities | |||||
| Previous MI | 10,085 (32.7%) | 2,784 (37.2%) | 2,672 (34.7%) | 2,417 (33.0%) | 2,211 (26.5%) |
| Hypertension | 19,718 (63.9%) | 4,285 (57.3%) | 4,617 (59.9%) | 4,855 (66.2%) | 5,961 (71.6%) |
| Diabetes | 7,743 (34.4%) | — | 2,034 (30.2%) | 2,499 (34.1%) | 3,188 (38.3%) |
| Stroke | 2,873 (9.3%) | 583 (7.8%) | 796 (10.3%) | 741 (10.1%) | 753 (9.0%) |
| Mean PREDICT score ‡ | 9.5 (9.4, 9.5) | 9.6 (9.5, 9.6) | 9.5 (9.4, 9.5) | 9.5 (9.4, 9.5) | 9.3 (9.2, 9.4) |
| Mean length of stay (days) | 7.9 (7.8, 8.1) | 10.1 (9.8, 10.5) | 8.4 (8.1, 8.6) | 7.0 (6.7, 7.3) | 6.4 (6.0, 6.8) |
| Median length of stay in (days) | 6.0 | 8.0 | 7.0 | 5.0 | 4.0 |
| Emergency medical services transport | 13,048 (42.3%) | 2,765 (36.9%) | 3,171 (41.2%) | 3,093 (42.1%) | 4,019 (48.2%) |
| Prehospital delay § <2 hours | 8,494 (27.4%) | 2,102 (27.9%) | 2,176 (28.2%) | 2,085 (28.4%) | 2,121 (25.4%) |
| Unknown | 3,632 (11.7%) | 912 (12.1%) | 857 (11.1%) | 821 (11.1%) | 1,043 (12.5%) |
| Event classification || | |||||
| STEMI | 6,106 (19.7%) | 1,474 (19.6%) | 1,926 (24.9%) | 1,420 (19.3%) | 1,284 (15.4%) |
| NSTEMI | 20,302 (65.5%) | 4,970 (66.1%) | 4,439 (57.4%) | 4,869 (66.0%) | 6,023 (72.1%) |
∗ Weighted number of definite or probable MI events.
† Chi-square test with Taylor series variance estimation for independence of characteristics across study years (categorical variables) or 1-way analysis of variance (continuous variables), significant for all variables at p <0.001.
‡ Modified PREDICT score did not include data on kidney function.
§ Prehospital delay was defined as the interval from earliest symptom onset time to hospital arrival time.
|| STEMI was defined as ST-segment elevation at any site on either the first or the last electrocardiogram.
Table 2 lists the proportion of patients receiving each medication and procedure of interest by year, age standardized to the 2000 United States census population. We observed increases in the use of all medications and procedures except for calcium channel blockers and thrombolytic therapy, which decreased throughout the study period. Data on the use of stents were first collected in 1998. Since then, the proportion of all patients with MIs receiving stents has doubled. Temporal trends for all patients and for those with STEMIs or NSTEMIs are illustrated for selected medications and procedures in Figure 1 .
| Therapy | Overall | 1987–1991 ‡ | 1992–1996 | 1997–2001 | 2002–2008 |
|---|---|---|---|---|---|
| (n = 30,986) † | (n = 7,524) | (n = 7,730) | (n = 7,380) | (n = 8,352) | |
| Medication | |||||
| Aspirin | 82.4% (0.22) | 65.7% (0.55) | 85.0% (0.41) | 89.0% (0.36) | 89.3% (0.34) |
| β blockers | 68.0% (0.27) | 45.3% (0.57) | 61.3% (0.55) | 75.5% (0.50) | 87.9% (0.36) |
| Calcium channel blockers | 44.0% (0.28) | 65.3 (0.55) | 54.5% (0.57) | 33.0% (0.55) | 24.8% (0.47) |
| ACE inhibitors | 40.9% (0.28) | — | 33.1% (0.54) | 58.7% (0.57) | 68.9% (0.51) |
| Heparin | 51.6% (0.28) | — | 62.4% (0.55) | 71.2% (0.53) | 70.3% (0.50) |
| Lipid-lowering medications | 29.0% (0.26) | — | — | 40.0% (0.57) | 72.2% (0.49) |
| Antiplatelet agents other than aspirin | 27.9% (0.25) | — | — | 45.5% (0.58) | 62.5% (0.53) |
| Procedures | |||||
| Thrombolytic agents | 11.7% (0.18) | 16.3% (0.43) | 18.5% (0.44) | 10.9% (0.36) | 2.0% (0.15) |
| PCI | 28.0% (0.25) | 15.6% (0.42) | 25.0% (0.49) | 30.4% (0.54) | 39.7% (0.54) |
| Stent | 14.4% (0.20) | — | — | 19.6% (0.46) | 35.9% (0.53) |
| CABG | 14.3% (0.20) | 15.6% (0.42) | 18.1% (0.44) | 14.5% (0.41) | 9.5% (0.32) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
