[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.163.94.5. Please contact the publisher to request reinstatement.
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Download PDF
Figure 1. Population Hierarchy in Applied Algorithm of Inclusion and Exclusion Criteria
Image description not available.
NRMI indicates National Registry of Myocardial Infarction.
Figure 2. Comparison of In-Hospital Mortality Rate by IMR Patient Status
Image description not available.
IMR indica tes immediate mechanical reperfusion.
Table 1. Contraindications for Thrombolytic Therapy by IMR Treatment Status*
Image description not available.
Table 2. Selected Characteristics of Patients With ST-Elevation Myocardial Infarction Who Had Contraindications for Thrombolytic Therapy*
Image description not available.
Table 3. Selected Patient Characteristics in Propensity Matched Analysis*
Image description not available.
Table 4. Association Between Selected Medications Administered Within 24 Hours of Presentation and IMR Treatment Status for Matched Patient Populations
Image description not available.
Table 5. Mortality Rates and In-Hospital Mortality Among Patients With ST-Elevation Myocardial Infarction Who Have Contraindications for Thrombolytic Therapy Presenting to All and IMR Capable Hospitals
Image description not available.
1.
Barron HV, Rundle AC, Gore JM.  et al. Participants in the National Registry of Myocardial Infarction-2.  Intracranial hemorrhage rates and effect of immediate beta-blocker use in patients with acute myocardial infarction treated with tissue plasminogen activator.  Am J Cardiol.2000;85:294-298.PubMed
2.
Every NR, Frederick PD, Robinson M.  et al.  A comparison of the National Registry of Myocardial Infarction 2 with the cooperative cardiovascular project.  J Am Coll Cardiol.1999;33:1886-1894.PubMed
3.
Canto JG, Every NR, Magid DJ.  et al. for the National Registry of Myocardial Infarction 2 Investigators.  The volume of primary angioplasty procedures and survival after acute myocardial infarction.  N Engl J Med.2000;342:1573-1580.PubMed
4.
Ryan TJ, Antman EM, Brooks NH.  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;34:890-911.PubMed
5.
Keeley E, Boura J, Grines C. Primary angioplasty versus intravenous thrombolytic therapy for acute myocardial infarction: a quantitative review of 23 randomized trials.  Lancet.2003;361:13-20.PubMed
6.
Magid D, Calonge B, Rumsfeld J.  et al.  Relation between hospital primary angioplasty volume and mortality for patients with acute MI treated with primary angioplasty vs thrombolytic therapy.  JAMA.2000;284:3131-3138.PubMed
7.
Himbert D, Juliard JM, Steg PG.  et al.  Primary coronary angioplasty for acute myocardial infarction with contraindication to thrombolysis.  Am J Cardiol.1993;71:377-381.PubMed
8.
McCullough PA, O'Neill WW, Graham M.  et al.  A prospective randomized trial of triage angiography in acute coronary syndromes ineligible for thrombolytic therapy: results of the Medicine versus Angiography in Thrombolytic Exclusion (MATE) trial.  J Am Coll Cardiol.1998;32:596-605.PubMed
9.
Zahn R, Schuster S, Schiele R.  et al. for the Maximal Individual Therapy in Acute Myocardial Infarction (MITRA) Study Group.  Comparison of primary angioplasty with conservative therapy in patients with acute myocardial infarction and contraindications for thrombolytic therapy.  Catheter Cardiovasc Interv.1999;46:127-133.PubMed
10.
Cragg DR, Friedman HZ, Bonema JD.  et al.  Outcome of patients with acute myocardial infarction who are ineligible for thrombolytic therapy.  Ann Intern Med.1991;115:173-177.PubMed
11.
Rosenbaum P, Rubin D. The central role of the propensity score in observational studies for causal effects.  Biometrika.1983;70:41-55.
12.
Rubin DB. Estimating causal effects from large data sets using propensity scores.  Ann Intern Med.1997;127(8 pt 2):757-763.PubMed
13.
Joffe MM, Rosenbaum PR. Invited commentary: propensity scores.  Am J Epidemiol.1999;150:327-333.PubMed
14.
Rogers WJ, Canto JG, Lambrew CT.  et al.  Temporal trends in the treatment of over 1.5 million patients with myocardial infarction in the US from 1990 through 1999: the National Registry of Myocardial Infarction 1,2, and 3.  J Am Coll Cardiol.2000;36:2056-2063.PubMed
15.
Rogers WJ, Bowlby LJ, Chandra NC.  et al.  Treatment of myocardial infarction in the United States (1990 to 1993): observations from the National Registry of Myocardial Infarction.  Circulation.1994;90:2103-2114.PubMed
16.
Peterson ED, Pollack CV, Roe MT.  et al.  Early use of glycoprotein IIb/IIIa inhibitors in non-ST-elevation acute myocardial infarction: observations from the National Registry of Myocardial Infarction 4.  J Am Coll Cardiol.2003;42:45-53.PubMed
17.
Parsons LS. Reducing bias in a propensity score matched-pair sample using greedy matching techniques. In: Proceedings of the Twenty-Sixth Annual SAS (Users Group International Conference), Cary, NC: SAS Institute Inc; 2001. Available at http://www2.sas.com/proceedings/sugi26/p214-26.pdf. Accessibility verified September 4, 2003.
18.
D'Agostino RB. Tutorial in biostatistics: propensity score medthods for bias reduction in the comparison of a treatment to a non-randomized control group.  Stat Med.1998;17:2265-2281.PubMed
19.
Stone GW, Grines CL, Browne KF.  et al. for the Primary Angioplasty in Myocardial Infarction (PAMI) Investigators.  Outcome of different reperfusion strategies in patients with former contraindications to thrombolytic therapy: a comparison of primary angioplasty and tissue plasminogen activator.  Cathet Cardiovasc Diagn.1996;39:333-339.PubMed
20.
Normand ST, Landrum MB, Guadagnoli E.  et al.  Validating recommendations for coronary angiography following acute myocardial infarction in the elderly: a matched analysis using propensity scores.  J Clin Epidemiol.2001;54:387-398.PubMed
21.
Moses LE. Measuring effects without randomization trials? options, problems, challenges.  Med Care.1995;33(suppl 4):AS8-AS14.PubMed
22.
Grines C, Westerhausen D, Grines L.  et al.  A randomized trial of transfer for primary angioplasty vs on-site thrombolysis in patients with high-risk myocardial infarction: the Air Primary Angioplasty in Myocardial Infarction study.  J Am Coll Cardiol.2002;39:1713-1719.PubMed
23.
Cannon C, Gibson C, Lambrew C.  et al.  Relationship of symptom-onset-to-balloon time and door-to-balloon time with mortality in patients undergoing angioplasty for acute myocardial infarction.  JAMA.2000;283:2941-2947.PubMed
24.
Aversano T, Aversano LT, Passami E.  et al.  Thrombolytic therapy vs primary percutaneous coronary intervention for myocardial infarction in patients presenting to hospitals without on-site cardiac surgery: a randomized controlled trial.  JAMA.2002;287:1943-1951.PubMed
25.
Widimsky P, Budesinsky T, Vorac D.  et al.  Long distance transport for primary angioplasty vs immediate thrombolysis in acute myocardial infarction: final results of the randomized national multicentre trial—PRAGUE-2.  Eur Heart J.2003;24:94-104.PubMed
26.
Moon JC, Kalra PR, Coats AJS. DANAMI-2: is primary angioplasty superior to thrombolysis in acute MI when the patient has to be transferred to an invasive centre?  Int J Cardiol.2002;85:199-201.PubMed
Original Contribution
October 8, 2003

Mortality Benefit of Immediate Revascularization of Acute ST-Segment Elevation Myocardial Infarction in Patients With Contraindications to Thrombolytic TherapyA Propensity Analysis

Author Affiliations

Author Affiliations: Department of Emergency Medicine (Drs Grzybowski, Clements, Welch, Ross, Tintinalli, and Zalenski), Center for Healthcare Effectiveness (Dr Grzybowski), and Department of Medicine, Division of Cardiology (Dr Zalenski), Wayne State University School of Medicine, Detroit, Mich; Department of Pharmacy, Spectrum-Health Hospital, Grand Rapids, Mich (Dr Clements), Ovation Research Group, Chicago, Ill (Ms Parsons), Department of Emergency Medicine, William Beaumont Hospital, Royal Oak, Mich (Dr Ross), and Section of Urgent Care, Department of Medicine, John D. Dingell Veterans Affairs Hospital, Detroit, Mich (Dr Zalenski).

JAMA. 2003;290(14):1891-1898. doi:10.1001/jama.290.14.1891
Context

Context There are no definitive recommendations for the management of acute myocardial infarction (AMI) in patients with ST-segment elevation who have contraindications to thrombolytic therapy. It is not clear whether, and the extent to which, immediate mechanical reperfusion (IMR) reduces in-hospital mortality in this population.

Objective To determine whether IMR (defined as percutaneous coronary intervention or coronary artery bypass graft surgery) is associated with a mortality benefit in patients with acute ST-segment elevation AMI who are eligible for IMR but have contraindications to thrombolytic therapy.

Design, Setting, and Patients From June 1994 to January 2003, the National Registry of Myocardial Infarction 2, 3, and 4 enrolled 1 799 704 patients with AMI. A total of 19 917 patients with acute ST-segment elevation were eligible for IMR but had thrombolytic contraindications after excluding patients who were transferred in from or out to other facilities, patients who received intracoronary thrombolytics, and those who received no medications within 24 hours of arrival.

Main Outcome Measure In-hospital mortality.

Results Of the 19 917 patients, 4705 patients (23.6%) received IMR and 5173 patients (25.9%) died. In-hospital mortality rates in the IMR and non-IMR treated groups in the unadjusted analysis were 11.1%, representing 521 of 4705 patients, and 30.6%, representing 4652 of 15 212 patients, respectively, for a risk reduction of 63.7% (odds ratio [OR], 0.28; 95% confidence interval [CI], 0.26-0.31). In a further analysis using a propensity matching score to reduce the effects of bias, 3905 patients who received IMR remained at lower risk for in-hospital mortality than 3905 matched patients (10.9% vs 20.1%, respectively, for a risk reduction of 45.8%; OR, 0.48; 95% CI, 0.43-0.55). Following a second logistic model applied to the matched groups to adjust for residual differences, a significant treatment effect persisted (OR, 0.64; 95% CI, 0.56-0.75).

Conclusions In this population, IMR was associated with a reduced risk of in-hospital mortality after appropriate adjustments. Of those we studied who were eligible for IMR, 15 212 patients (76.4%) did not receive it. These results suggest that using IMR in patients with acute ST-segment elevation AMI and contraindications to thrombolytics should be strongly considered.

Thrombolytic therapy is the most widely used method of achieving coronary artery reperfusion in patients with acute ST-segment elevation myocardial infarction (STEMI).1,2 Evidence shows that percutaneous coronary intervention may be more effective than thrombolytic therapy, but this treatment is limited to institutions capable of 24-hour cardiac catheterization, percutaneous coronary intervention, and coronary artery bypass graft surgery.36 The 1999 American College of Cardiology/American Heart Association (ACC/AHA) guidelines classified mechanical reperfusion as a class IIA recommendation (conflicting evidence but weight in favor of efficacy) for managing acute myocardial infarction (AMI) in patients with STEMI who have contraindications to thrombolytic therapy.4

It is not clear whether immediate mechanical reperfusion (IMR) among patients who have contraindications for thrombolytic therapy improves outcome. It has been reported that the in-hospital mortality rate among patients with STEMI who have contraindications to thrombolytic therapy and treated with primary angioplasty is 9.0%.7 However, to our knowledge, there are no data from trials comparing primary angioplasty to conservative treatment therapy among the STEMI patient population who have contraindications to thrombolytic therapy. Previous comparison trials were not limited to STEMI patients or a 12-hour reperfusion window and included patients who did not meet criteria for thrombolysis even though they may or may not have had contraindications to thrombolytic therapy.810

The objective of this study was to determine whether receiving IMR was associated with a mortality benefit in patients with acute STEMI who were eligible for IMR and had contraindications to thrombolytic therapy. Because confounding factors and selection bias may affect the validity of observational studies measuring treatment effects, we performed a propensity analysis to answer this question.1113

METHODS
Data Source

The National Registry of Myocardial Infarction (NRMI) is a large voluntary observational database of hospitalized patients with confirmed AMI diagnosis that began in 1990. Patients included in the present study were derived from the NRMI 2, 3, and 4. The NRMI 2 enrolled patients at 1674 hospitals from June 1994 through April 1998; NRMI 3 enrolled patients at 1553 hospitals from May 1998 through June 2000; and NRMI 4 enrolled patients at 1273 hospitals from July 2000 through January 2003.

The diagnosis of myocardial infarction was based on patient history and presentation suggestive of AMI accompanied by at least one of the following: (1) a total creatine kinase or creatine kinase-MB greater than or equal to twice the upper limit of normal (normal as defined by an individual hospital's laboratory definition); (2) electrocardiographic (ECG) evidence of AMI; (3) elevation of cardiac specific enzymes (NRMI 3 and 4), scintigraphic or autopsy evidence of AMI, or discharge diagnosis of AMI (International Classification of Diseases, Ninth Revision, Clinical Modification code 410.X1 for NRMI 2). For NRMI 3 and 4, a discharge diagnosis was required as was the above clinical criteria.

All study site coordinators were formally trained and given a manual of operations to correctly abstract the data from the medical records and transcribe the data onto case report forms. The case report forms were then forwarded to StatProbe, Inc (Lexington, Ky), where the data underwent double-data entry or electronic scanning and multiple electronic checks for accuracy. Any unrecorded fields or inconsistencies were flagged and returned to the site for clarification and correction. The NRMI collection procedures have been previously reported.1416 The NRMI's data set has been validated by comparing it with the Cooperative Cardiovascular Project even though the NRMI includes all payers instead of only Medicare beneficiaries.2

Study Population

All patients in the NRMI 2, 3, and 4 databases were eligible for study inclusion (n = 1 799 704). An algorithm was used to identify the desired study population (Figure 1). Patients were included if (1) the initial ECG result demonstrated criteria for thrombolytic therapy defined as STEMI or presumed new left bundle-branch block, (2) their symptom onset to hospital door time was less than 12 hours, and (3) they had a contraindication to receive intravenous thrombolytic therapy, defined as the presence of one or more of the following: active internal bleeding or known bleeding diathesis, a history of cerebrovascular accident, severe uncontrolled hypertension, recent surgery or trauma, or intracranial neoplasm, arteriovenous malformation, or aneurysm or traumatic cardiopulmonary resuscitation. Due to changes in the NRMI 4 data collection procedures, contraindications to thrombolytics among nonreperfused patients were derived as stated above, but for reperfused patients, contraindications to thrombolytics were determined via medical history of stroke or bleeding diathesis. Exclusion criteria consisted of patients who (1) were either transferred in from or out to other facilities (due to incomplete data), (2) received intracoronary thrombolytic therapy, or (3) had no documentation of any medication given in the first 24 hours of hospitalization (because these cases may have received nonstandard treatment). A total of 19 917 patients eligible to receive IMR treatment were included in the final study population.

Study Variables

Registry variables collected included demographics, previous medical history, presentation data, medications administered within 24 hours of presentation, treatment, procedures, and in-hospital mortality. Time from symptom onset to hospital arrival was abstracted from the medical record. To prevent the influence of outliers, any patient with a value greater than 72 hours was assigned a time of 72 hours. The following variable values were assigned to missing: heart rates less than 1/min or greater than 220/min, systolic blood pressures less than 1 or greater than 250 mm Hg, and diastolic blood pressures less than 1 or greater than 130 mm Hg. The outcome variable was in-hospital mortality.

Statistical Analysis

Patients were divided into 2 groups—those who did and those who did not receive IMR treatment. Immediate mechanical reperfusion treatment was defined as having received percutaneous coronary intervention or coronary artery bypass graft surgery within 12 hours of symptom onset.4

Baseline differences between those receiving and not receiving IMR were compared using either the t test or the Wilcoxon rank-sum test for continuous variables and the χ2 statistic for categorical variables. Ninety-five percent confidence intervals (CIs) were calculated for point estimates. Unadjusted odds ratios (ORs) and their 95% CIs were calculated for each predictor variable and IMR status.

Because receiving IMR was not randomly assigned in this study population, propensity scores were calculated to account for potentially confounding factors and selection biases.1113 Using multivariate logistic regression, the propensity for receiving IMR was determined. Predictor variables considered for the propensity derivation model included demographics, prior medical history, and presentation data. The final derivation model included the following predictor variables: age; sex; race; prehospital ECG status; symptom onset to door time; a medical history of smoking, diabetes, previous myocardial infarction, hypercholesterolemia, congestive heart failure, hypertension, percutaneous coronary intervention, coronary artery bypass graft surgery, angina, stroke, renal disease, chronic obstructive pulmonary disease, and family history of coronary artery disease; presentation characteristics including chest pain at presentation, cardiogenic shock, systolic and diastolic blood pressures, heart rate, and year of discharge. The c statistic for the propensity score derivation model was 0.80, indicating a strong ability to discriminate between patients who received and who did not receive IMR. A propensity score for IMR treatment was then calculated for each patient from the logistic equation. The propensity score for all matched patients (n = 7810) ranged from 0.0011034 to 0.9285301, which in effect, reflected the probability that a patient would receive IMR treatment.

The propensity scores were used to individually match patients receiving IMR to unique control patients using a SAS macro.17 To match each IMR patient to a non-IMR patient, we used a propensity score that was identical to 8 digits. If the score could not be matched up to 8 digits, a patient receiving IMR was excluded. We were successfully able to match 3905 patients receiving IMR to 3905 unique patients not receiving IMR. In-hospital mortality rates by IMR status were compared within the propensity matched sample.

Following the identification of the matched sample, residual imbalances were adjusted for in the analysis by a second multivariate logistic regression model applied to the matched groups. The purpose of this model was to derive the adjusted ORs for in-hospital mortality and IMR treatment after controlling for medications using the generalized estimating equation in PROC GENMOD. To determine if results were similar to a more conventional approach to remove bias, we performed a supplementary analysis of mortality rates based on a propensity score quintile approach.18

To determine if the mortality reduction differed in patients presenting to hospitals capable of IMR, we repeated the analyses in this group of patients using the same inclusion and exclusion criteria (n = 17 418). In this population, 3756 IMR patients were matched to 3756 non-IMR patients. Finally, we conducted a sensitivity analysis and modeled the potential benefit of IMR after broadening our inclusion criteria to those who were transferred to a NRMI hospital, received no medications within 24 hours, and received intracoronary thrombolytic therapy (n = 26 263). Patients who received intracoronary thrombolytics were classified as having had IMR, and 5320 IMR patients were matched to 5320 non-IMR patients.

All reported P values are 2-sided with a significance level of .05 for detecting a difference. Appropriate regression diagnostics were performed to confirm the validity of the multivariate models, including calculating the goodness-of-fit test statistic, testing for outliers, and assessing the predicted outcomes by actual outcomes via classification tables. All analyses were performed using SAS software, version 8.2 (SAS Institute, Cary, NC).

RESULTS
Patient Characteristics

The final study population eligible for IMR with contraindications to thrombolytics included 19 917 acute STEMI patients (1.1%). Of these, 4705 patients (23.6%; 95% CI, 23.0%-24.2%) received IMR. Overall, 5173 patients (25.9%) died during their hospital stay. Table 1 depicts the prevalence of thrombolytic contraindications and their associations with IMR status. Contraindications included a history of cerebrovascular accident (29.9%) followed by recent trauma or surgery (22.1%), active internal bleeding (18.1%), traumatic cardiopulmonary resuscitation (16.4%), severe, uncontrolled hypertension (7.4%), and intracranial neoplasm (3.6%). We found that patients with cerebrovascular accidents were less likely to receive IMR. Those patients with recent surgery or trauma, severe or uncontrolled hypertension, or traumatic cardiopulmonary resuscitation were more likely to receive IMR than those without these contraindications. The majority of patients (62.7%) had only 1 contraindication to thrombolytic therapy.

Baseline demographics, past medical history, and presentation characteristics according to IMR treatment are summarized in Table 2. Patients receiving IMR were younger and more likely to be men; they were more likely to have a history of smoking, hypercholesterolemia, coronary artery disease, previous stroke, and percutaneous coronary intervention. Patients treated with IMR were also more likely to present with cardiogenic shock, chest pain, and have prehospital ECGs. Lastly, they were more likely to receive aspirin, other antiplatelet agents, heparin, other antithrombins, β-blockers, intravenous nitroglycerin, and lidocaine. Upon presentation, patients who were not treated with IMR had significantly higher heart rates (P<.001), and slightly higher systolic blood pressures (P<.01), and symptom onset to door times (P<.001). Additionally, patients not receiving IMR were more likely to have a medical history of renal insufficiency, diabetes mellitus, myocardial infarction, hypertension, congestive heart failure, coronary artery bypass surgery, chronic obstructive pulmonary disease, and angina. Non-IMR treated patients were also more likely to receive calcium channel blockers.

Patient Characteristics Following Matched Propensity Score Analysis

Baseline demographics, past medical history, and presenting characteristics comparing the IMR and non-IMR, treatment-matched patients are shown in Table 3. Of 4705 patients who received IMR, 3905 (83%) were successfully matched to non-IMR patients. As shown in Table 3, the propensity score matching technique eliminated all significant differences between measured patient demographics, past medical history, and presentation data by IMR treatment status. Table 4 shows that significant differences by IMR treatment status remained similar to medications administered within 24 hours of emergency department presentation. The IMR treated patients were still more likely to have been treated with aspirin, other antiplatelet agents, heparin, other antithrombin agents, intravenous nitroglycerin, and lidocaine. With the exception of other antithrombin agents and lidocaine, the magnitude of the associations between aspirin, other antiplatelets, and intravenous nitroglycerin and receiving IMR were reduced following the matched analysis. The propensity analysis only removed the significant effects of β-blockers and calcium channel blockers.

IMR Treatment Use and Mortality

Table 5 summarizes in-hospital mortality rates and odds of in-hospital mortality by statistical model for patients presenting to all NRMI hospitals, IMR-capable NRMI hospitals, and for patients in our sensitivity analysis presenting to all NRMI hospitals. Immediate mechanical reperfusion use was statistically significantly associated with reduced mortality in all unadjusted and adjusted analyses, P<.001. For all study populations, the reduction in mortality decreased as more adjustments were made to the model. Figure 2 shows in-hospital mortality rates by propensity score quintile for IMR status using a more conventional propensity score approach. Except for the first quintile, in-hospital mortality rates were still significantly lower in patients who received IMR than in those that did not: 0.84 (95% CI, 0.55-1.26; P = .40) for quintile 1; 0.62 (95% CI, 0.48-0.80; P<.001) for quintile 2; 0.50 (95% CI, 0.40-0.62; P<.001) for quintile 3; 0.24 (95% CI, 0.20-0.30; P<.001) for quintile 4; and 0.30 (95% CI, 0.23-0.39; P<.001) for quintile 5.

COMMENT

In this study, patients with acute STEMI and contraindications to thrombolytic therapy had a substantial in-hospital mortality rate of 25.9%. In the unmatched and matched analyses, the mortality reductions associated with IMR treatment were 63.7% and 45.8%, respectively. In addition, after adjusting for medications in the matched patient sample, IMR had a protective effect against in-hospital mortality. These results corroborate results from the analysis examining in-hospital mortality rates by propensity score quintile. Our finding that only 23.6% of the overall study group received IMR suggests that there is potential opportunity to improve quality of care.

This study is unique because it analyzes STEMI patients with contraindications to thrombolytic therapy who are eligible for reperfusion and because it quantifies the benefit of IMR treatment in such patients. These findings are consistent with previous studies. Himbert et al7 found a 9.0% mortality rate in a case series of 45 patients with contraindications to thrombolytic therapy treated within 6 hours with percutaneous coronary intervention. The analysis performed by Stone et al19 of a subgroup of 151 patients in the Primary Angioplasty in Myocardial Infarction trial that was considered to have an unfavorable risk/benefit ratio (age >70 years, symptom onset >4 hours, or prior coronary artery bypass graft surgery) found that IMR treatment was associated with an absolute mortality reduction of 10.3% over treatment with thrombolytic therapy. This benefit of IMR over thrombolytic therapy in this subgroup suggests that such higher risk patients may also be at the higher end of the spectrum of the relative benefit of IMR compared with thrombolytic treatment. In the Maximal Individual Therapy in Acute Myocardial Infarction study9 of 337 patients with 1 or more contraindications to thrombolytic therapy who presented within 96 hours of symptoms onset, the 13.6% of patients treated with percutaneous coronary intervention had a 54% reduction in the proportion of in-hospital deaths compared with nonreperfused patients in a model adjusting for covariates.19

This study has several notable strengths. First, our study overcomes the limitations of previous studies by including a large population of thrombolytic contraindicated patients after applying vigorously defined criteria for reperfusion therapy. Second, unlike previous observational studies of IMR treatment and mortality, our sample size afforded us the ability to conduct a propensity analysis, which is an effective method in reducing bias in baseline characteristics when assessing treatment effects.11,20 Third, our data are drawn from a wide variety of hospitals and may have greater external validity compared with those enrolled in randomized trials.21 Finally, after broadening our criteria for the sensitivity analysis, our unadjusted and adjusted sensitivity analyses that modeled the benefit of IMR showed very similar mortality reductions thereby strengthening our study since it could be argued that our predefined criteria may have driven the decision to proceed with IMR.

Our primary limitation is that our data are not from a prospective randomized trial; therefore, we could adjust for only the observed confounding variables. It is important to emphasize that reductions in mortality for the IMR group may reflect other unmeasured factors, such as better prehospital care, presenting symptoms,3,9 or other medical care.9 For example, in our unmatched population, we found that IMR patients were, on average, 7.4 years younger, 1.92 times more likely to have prehospital ECGs, 5.10 times more likely to present with chest pain, and more likely to be given certain cardiac medications. After matching, we found that patients treated by IMR were still more likely to receive aspirin, other antiplatelet medications, heparin, and other antithrombin agents, which suggests that IMR-treated patients may have been treated more aggressively but in ways that could not have been measured and controlled for analytically. We did attempt to control for hospital-specific factors by performing the analyses among patients presenting to hospitals capable of percutaneous coronary interventions.

Although the NRMI registry includes a broad spectrum of hospitals across the country, differences in hospitals that did and did not participate may exist thereby generating unmeasured bias. Compared with nonparticipating NRMI hospitals, NRMI participating hospitals tend to be slightly larger and have more tertiary cardiac facilities.16 The inability to determine who, among patients with contraindications to thrombolytic therapy, is most likely to benefit from IMR treatment further limits our study. Lastly, because NRMI categorically collects medication and therapy data administered within 24 hours after arrival in a hospital, we are unable to discern the effect of appropriately timed therapy relative to risk of mortality.

Our results suggest that it is important to consider methods for improving utilization of and access to IMR. One possibility is to transport patients to IMR-capable facilities from either the incident site or a non–IMR-capable hospital. However, pilot studies have demonstrated delays in treatments when patients are transferred, and this holds true when patients are randomized on-site to receive thrombolysis treatment or are transported to an IMR-capable hospital, but these data are limited to patients with no contraindications to thrombolytic therapy.22,23

Various approaches to providing timely IMR to STEMI patients have been studied.24,25 In Europe, prehospital diversion of STEMI patients to IMR facilities has been shown to reduce the incidence of the combined clinical end point of death, reinfarction, or stroke at 30 days compared with thrombolysis treatment at community hospitals in both the PRimary Angioplasty in patients transferred from General community hospitals to specialized PTCA Units with or without Emergency thrombolysis (PRAGUE-2)25 (8.4% vs 15.2%) and the DANish trial in Acute Myocardial Infarction-226 (8.0% vs 13.7%) studies.25,26 Results from these studies have prompted a call for specialized centers for STEMI patients, similar to the way most trauma systems operate.

In conclusion, our results demonstrated a mortality reduction if IMR is used for patients with STEMI who have contraindications to thrombolytic therapy. Future studies are needed to confirm our results and to identify optimal methods of providing rapid access to IMR in this population.

References
1.
Barron HV, Rundle AC, Gore JM.  et al. Participants in the National Registry of Myocardial Infarction-2.  Intracranial hemorrhage rates and effect of immediate beta-blocker use in patients with acute myocardial infarction treated with tissue plasminogen activator.  Am J Cardiol.2000;85:294-298.PubMed
2.
Every NR, Frederick PD, Robinson M.  et al.  A comparison of the National Registry of Myocardial Infarction 2 with the cooperative cardiovascular project.  J Am Coll Cardiol.1999;33:1886-1894.PubMed
3.
Canto JG, Every NR, Magid DJ.  et al. for the National Registry of Myocardial Infarction 2 Investigators.  The volume of primary angioplasty procedures and survival after acute myocardial infarction.  N Engl J Med.2000;342:1573-1580.PubMed
4.
Ryan TJ, Antman EM, Brooks NH.  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;34:890-911.PubMed
5.
Keeley E, Boura J, Grines C. Primary angioplasty versus intravenous thrombolytic therapy for acute myocardial infarction: a quantitative review of 23 randomized trials.  Lancet.2003;361:13-20.PubMed
6.
Magid D, Calonge B, Rumsfeld J.  et al.  Relation between hospital primary angioplasty volume and mortality for patients with acute MI treated with primary angioplasty vs thrombolytic therapy.  JAMA.2000;284:3131-3138.PubMed
7.
Himbert D, Juliard JM, Steg PG.  et al.  Primary coronary angioplasty for acute myocardial infarction with contraindication to thrombolysis.  Am J Cardiol.1993;71:377-381.PubMed
8.
McCullough PA, O'Neill WW, Graham M.  et al.  A prospective randomized trial of triage angiography in acute coronary syndromes ineligible for thrombolytic therapy: results of the Medicine versus Angiography in Thrombolytic Exclusion (MATE) trial.  J Am Coll Cardiol.1998;32:596-605.PubMed
9.
Zahn R, Schuster S, Schiele R.  et al. for the Maximal Individual Therapy in Acute Myocardial Infarction (MITRA) Study Group.  Comparison of primary angioplasty with conservative therapy in patients with acute myocardial infarction and contraindications for thrombolytic therapy.  Catheter Cardiovasc Interv.1999;46:127-133.PubMed
10.
Cragg DR, Friedman HZ, Bonema JD.  et al.  Outcome of patients with acute myocardial infarction who are ineligible for thrombolytic therapy.  Ann Intern Med.1991;115:173-177.PubMed
11.
Rosenbaum P, Rubin D. The central role of the propensity score in observational studies for causal effects.  Biometrika.1983;70:41-55.
12.
Rubin DB. Estimating causal effects from large data sets using propensity scores.  Ann Intern Med.1997;127(8 pt 2):757-763.PubMed
13.
Joffe MM, Rosenbaum PR. Invited commentary: propensity scores.  Am J Epidemiol.1999;150:327-333.PubMed
14.
Rogers WJ, Canto JG, Lambrew CT.  et al.  Temporal trends in the treatment of over 1.5 million patients with myocardial infarction in the US from 1990 through 1999: the National Registry of Myocardial Infarction 1,2, and 3.  J Am Coll Cardiol.2000;36:2056-2063.PubMed
15.
Rogers WJ, Bowlby LJ, Chandra NC.  et al.  Treatment of myocardial infarction in the United States (1990 to 1993): observations from the National Registry of Myocardial Infarction.  Circulation.1994;90:2103-2114.PubMed
16.
Peterson ED, Pollack CV, Roe MT.  et al.  Early use of glycoprotein IIb/IIIa inhibitors in non-ST-elevation acute myocardial infarction: observations from the National Registry of Myocardial Infarction 4.  J Am Coll Cardiol.2003;42:45-53.PubMed
17.
Parsons LS. Reducing bias in a propensity score matched-pair sample using greedy matching techniques. In: Proceedings of the Twenty-Sixth Annual SAS (Users Group International Conference), Cary, NC: SAS Institute Inc; 2001. Available at http://www2.sas.com/proceedings/sugi26/p214-26.pdf. Accessibility verified September 4, 2003.
18.
D'Agostino RB. Tutorial in biostatistics: propensity score medthods for bias reduction in the comparison of a treatment to a non-randomized control group.  Stat Med.1998;17:2265-2281.PubMed
19.
Stone GW, Grines CL, Browne KF.  et al. for the Primary Angioplasty in Myocardial Infarction (PAMI) Investigators.  Outcome of different reperfusion strategies in patients with former contraindications to thrombolytic therapy: a comparison of primary angioplasty and tissue plasminogen activator.  Cathet Cardiovasc Diagn.1996;39:333-339.PubMed
20.
Normand ST, Landrum MB, Guadagnoli E.  et al.  Validating recommendations for coronary angiography following acute myocardial infarction in the elderly: a matched analysis using propensity scores.  J Clin Epidemiol.2001;54:387-398.PubMed
21.
Moses LE. Measuring effects without randomization trials? options, problems, challenges.  Med Care.1995;33(suppl 4):AS8-AS14.PubMed
22.
Grines C, Westerhausen D, Grines L.  et al.  A randomized trial of transfer for primary angioplasty vs on-site thrombolysis in patients with high-risk myocardial infarction: the Air Primary Angioplasty in Myocardial Infarction study.  J Am Coll Cardiol.2002;39:1713-1719.PubMed
23.
Cannon C, Gibson C, Lambrew C.  et al.  Relationship of symptom-onset-to-balloon time and door-to-balloon time with mortality in patients undergoing angioplasty for acute myocardial infarction.  JAMA.2000;283:2941-2947.PubMed
24.
Aversano T, Aversano LT, Passami E.  et al.  Thrombolytic therapy vs primary percutaneous coronary intervention for myocardial infarction in patients presenting to hospitals without on-site cardiac surgery: a randomized controlled trial.  JAMA.2002;287:1943-1951.PubMed
25.
Widimsky P, Budesinsky T, Vorac D.  et al.  Long distance transport for primary angioplasty vs immediate thrombolysis in acute myocardial infarction: final results of the randomized national multicentre trial—PRAGUE-2.  Eur Heart J.2003;24:94-104.PubMed
26.
Moon JC, Kalra PR, Coats AJS. DANAMI-2: is primary angioplasty superior to thrombolysis in acute MI when the patient has to be transferred to an invasive centre?  Int J Cardiol.2002;85:199-201.PubMed
×