Outcomes Among Patients With ST-Segment–Elevation Myocardial Infarction Presenting to Interventional Hospitals With and Without On-Site Cardiac Surgery
Background— Primary percutaneous coronary intervention (pPCI) is the preferred reperfusion strategy for patients with ST-segment–elevation myocardial infarction (STEMI). The quality of care and safety and efficacy of pPCI at hospitals without on-site open heart surgery (No-OHS hospitals) remains an area of active investigation.
Methods and Results— The National Registry of Myocardial Infarction enrolled 58 821 STEMI patients from 214 OHS hospitals (n=54 076) and 52 No-OHS hospitals (n=4745) with PCI capabilities from 2004 to 2006. Patients presenting to OHS hospitals had substantially lower in-hospital mortality (7.0% versus 9.8%, P<0.001) and were more likely to receive any form of acute reperfusion therapy (80.8% versus 70.8%, P<0.001). Patients who presented to OHS hospitals were more likely to receive guideline recommended medications within 24 hours of arrival. In a propensity score model matching for patient characteristics and transfer status, in-hospital mortality remained significantly lower among patients presenting to OHS hospitals (7.2% versus 9.3%, P=0.025). When this model was further adjusted for differences in the use of acute reperfusion therapy, medications administered within 24 hours and hospital characteristics, the mortality difference was of borderline significance (hazard ratio, 0.87; 95% CI, 0.75 to 1.01; P=0.067). When the propensity score analysis was restricted to patients who underwent pPCI, there was no significant difference in mortality (3.8% versus 3.3%, P=0.44).
Conclusions— STEMI patients presenting to No-OHS hospitals have substantially higher mortality, are less likely to receive guideline recommended medications within 24 hours, and are less likely to undergo acute reperfusion therapy, although this difference was of borderline significance after adjusting for hospital and treatment variables. There was no difference in mortality among patients undergoing pPCI.
Received December 8, 2008; accepted August 19, 2009.
The primary goal in the management of patients with ST-segment–elevation myocardial infarction (STEMI) is rapid and complete reperfusion of the infarct-related artery.1–3 Clinical trials have shown improved outcomes among STEMI patients undergoing primary percutaneous coronary intervention (pPCI) compared with fibrinolytic therapy,4–7 and pPCI has rapidly become the preferred initial reperfusion strategy in the United States and internationally.8,9
Through the development of better interventional cardiology procedures and equipment, complications from PCI that necessitate emergent cardiac surgery are uncommon although unpredictable.10 Although there is evidence that pPCI performed at hospitals without on-site cardiac surgery (No-OHS hospitals) improves door-to-balloon time and is safe,11–13 other studies have reported higher mortality among patients undergoing both elective PCI14 and pPCI15 at No-OHS hospitals. Although these studies have flaws, there remains an ongoing debate as to the safety of pPCI at hospitals without back-up cardiac surgery.
A recent study from the National Cardiovascular Data Registry reported that in-hospital mortality was no different among 33 002 patients undergoing pPCI at No-OHS hospitals after adjustment for site characteristics and potential confounding variables.16
Whereas prior studies reported outcomes solely among patients undergoing pPCI, the goal of this analysis was to evaluate clinical outcomes and the use of guideline-recommended therapies among all patients presenting to hospitals with and without on-site cardiac surgery. A secondary goal was to evaluate outcomes specifically among patients undergoing pPCI at the 2 hospital types.
WHAT IS KNOWN
Among patients with ST-segment–elevation myocardial infarction, primary percutaneous coronary intervention is the preferred method of reperfusion therapy.
Whether the presence of on-site back-up cardiac surgery should be necessary for the support of cardiac catheterization laboratories is an area of active investigation, and prior studies evaluating the role of back-up cardiac surgery for primary percutaneous coronary intervention have presented mixed results.
Performing primary percutaneous coronary intervention at hospitals without back-up cardiac surgery likely improves door-to-balloon time but has been associated with a higher incidence of adverse cardiovascular outcomes.
WHAT THE STUDY ADDS
In this analysis of data from the National Registry of Myocardial Infarction, ST-segment–elevation myocardial infarction patients presenting to hospitals without back-up cardiac surgery had higher in-hospital mortality and were less likely to receive acute reperfusion therapy when compared with those who presented to hospitals with back-up cardiac surgery, even in a propensity-matched analysis matching for patient characteristics.
However, when this propensity-matched analysis was further adjusted for differences in the use of reperfusion therapy, medications within 24 hours, and other hospital characteristics, the difference in mortality was attenuated, suggesting that the difference in outcomes between hospitals with and without cardiac surgery is driven by factors other than the presence of on-site surgery.
The National Registry of Myocardial Infarction (NRMI) is an industry-sponsored observational study whose methods have previously been described.17,18 To be included in the registry, patients must have had an acute myocardial infarction (MI) documented according to local hospital criteria, usually including a history suggestive of acute MI and corroborated by cardiac biomarkers, 12-lead ECG, coronary angiography, or ICD-9 diagnostic code of MI. STEMI was defined as ST-segment elevation or left bundle branch block (old/new/unknown) on first/subsequent 12-lead ECG. Reperfusion-eligible patients were defined as those with STEMI and prehospital delay <12 hours for whom the first ECG was diagnostic.
Only hospitals with interventional capabilities were included in this analysis. A registry coordinator at each participating hospital recorded data from each patient, including demographic information and time intervals, including symptom onset-to-door time. In-hospital mortality was available among patients who were not transferred to another facility. Other adverse outcomes, including the composite of death and MI and the composite of death, MI, cardiogenic shock, and congestive heart failure (CHF), were assessed among patients who were not transferred.
To ensure quality control of registry data, registry coordinators were trained in data entry using a standardized manual of instructions and definitions. Case report forms were required to pass systematic range and internal consistency checking. Hospitals obtained approval of the registry data collection process as dictated by local investigational review boards.
All statistical analyses were performed using a commercially-available statistical package (SAS 9.1.3 Service Pack 4, SAS Institute). Continuous variable values are reported as the mean±SD. Propensity scores were developed to account for biases that may exist between patients who presented to hospitals with and without on-site cardiac surgery. For descriptive purposes, propensity score matching was performed. In multivariable models, the unmatched data were used with propensity score included as a covariate. For continuous and dichotomous variables, comparisons were assessed for matched and unmatched data using the standardized difference (d). To account for clustering of patients within hospitals, model-based group comparisons of unmatched binary, continuous, and ordinal data were implemented using a generalized estimating equations approach or generalized linear mixed models. To estimate group differences in survival, patients were considered to enter the risk set at their time of presentation (ie, left truncation was present in the sample). This accounts for potential survival bias associated with differential time to arrival. Cox proportional hazard regression models accounting for both left-truncated and right-censored data were constructed, with adjustment for propensity score and covariates. Treating the transfer-in patient as the truncating event and transfer-out, death, or discharge events as the censoring event, Cox regression analysis was applied to adjust for the administration of guideline-recommended medications (aspirin, clopidogrel, statins, β-blockers, and glycoprotein IIb/IIIa inhibitors) within 24 hours, the use of initial reperfusion therapy (pPCI, facilitated PCI, fibrinolytic administration or immediate coronary artery bypass grafting), and hospital characteristics (region in United States, hospital type, hospital size, and MI volume). Hospitals, rather than patients, were used as the sampling unit to account for the clustered nature of the data. The Thrombolysis in Myocardial Infarction Risk Score19 and Thrombolysis in Myocardial Infarction Risk Index20 were calculated only in patients for whom all the data were available. All probability values used 2-tailed tests, and probability values <0.05 were considered significant.
In the propensity-matched analysis, patients presenting to No-OHS hospitals were matched to patients presenting to OHS hospitals using the propensity score generated from a nonparsimonious logistic regression model using the outcome variable of No-OHS hospital (yes/no). No interactions between the predictor variables were used in the models. A greedy matching algorithm was used to match patients based on 8→1 digit matching. The following variables were used to generate the propensity score: transferred-in status, sex, age, hypertension, diabetes mellitus, hyperlipidemia, stroke, peripheral vascular disease, previous MI, CHF, smoking status, anterior infarct location and aspirin, clopidogrel, β-blocker, and statin received within 24 hours before arrival. Patients grouped into OHS hospitals were randomly selected without replacement.
There were 186 267 patients enrolled at 456 hospitals in NRMI 5 from April 2004 to December 2006. Of these, 158 892 STEMI and NSTEMI patients presented to 266 hospitals with pPCI capabilities. There were 58 821 patients with STEMI, of whom 4745 (8.1%) presented to 52 No-OHS hospitals and 54 076 (91.9%) presented to 214 OHS hospitals. Baseline characteristics among the entire STEMI population are shown in Table 1⇓. Because of the large size of the population, there were statistically significant differences in many baseline characteristics. Patients presenting to OHS hospitals were more likely male, had a higher prevalence of hyperlipidemia and a lower prevalence of stroke, atrial fibrillation, and heart failure, had lower pulse rate, systolic blood pressure, and creatinine concentration, and were more likely to have a Killip class >1. Patients presenting to OHS hospitals were considerably more likely to have been transferred from another hospital (39.6% versus 6.7%, P<0.001).
Patients presenting to OHS hospitals were significantly more likely to receive aspirin (91.9% versus 84.5%, P<0.001), clopidogrel (62.7% versus 52.9%, P=0.026), β-blockers (82.7% versus 76.1%, P<0.001), and statins (50.4% versus 38.3%, P<0.001) within the first 24 hours after arrival (Table 2).
Among the 39 131 reperfusion-eligible patients (66.5%), those presenting to OHS hospitals were more likely to receive any form of reperfusion therapy (80.8% versus 70.8%, P<0.001), and there was a strong trend toward more use of pPCI (55.6% versus 43.8%, P=0.057) at OHS hospitals (Table 3). Among those who did not undergo reperfusion therapy, the most common reasons cited at both types of hospitals were Do Not Resuscitate orders, patient/family refusal, and quality of life decisions.
Among nontransfer-in patients for whom time-to-reperfusion data were available, there was no significant difference in mean or median door-to-balloon time between the hospital types (Table 3). Mean and median door-to-needle times were longer at OHS hospitals. Patients presenting to OHS hospitals were more likely to undergo rescue PCI for presumed failed fibrinolytic therapy (5.9% versus 1.9%, P<0.001). Immediate CABG was performed in 2.0% of patients presenting to OHS hospitals.
Patients who presented to No-OHS hospitals were far more likely to be transferred to another short-term general hospital (36.5% versus 2.2%, P<0.001). Patients who were transferred out from No-OHS hospitals and those transferred in to OHS hospitals were relatively similar except that patients transferred from No-OHS hospitals were significantly less likely to have hyperlipidemia (36.5% versus 47.2%, P<0.001), chronic obstructive pulmonary disease (8.3% versus 11.1%, P<0.001), peripheral arterial disease (4.2% versus 6.0%, P=0.02), chronic kidney disease (2.4% versus 4.2%, P<0.001), and angina pectoris (6.0% versus 10.6%, P=0.03) and were more likely to have undergone CABG (8.5% versus 6.8%, P=0.01).
Among patients who were not transferred (n=55,903), in-hospital mortality was significantly lower among patients presenting to OHS hospitals (7.0% versus 9.8%, P=0.001) (Table 4). There was no significant difference in the incidence of recurrent MI (1.9% versus 1.5%, P=0.30), but the composite of death and MI (8.5% versus 10.9%, P=0.014) was lower among patients presenting to OHS hospitals. Among patients who underwent immediate CABG at OHS hospitals, the in-hospital mortality was 9.4%. Among patients who were transferred in to OHS hospitals, the in-hospital mortality was 5.9%.
Among patients who survived and were not transferred, there were small but significant differences favoring OHS hospitals in receipt of guideline-recommended medications at discharge, including aspirin (93.9% versus 89.9%, P<0.001), β-blockers (90.7% versus 87.9%, P=0.037), and statins (84.0% versus 77.0%, P=0.003; Table 5).
Patients presenting to No-OHS hospitals were matched 1:1 in a propensity-matched model with patients presenting to OHS hospitals based on the following baseline characteristics: transferred-in status, sex, age, hypertension, diabetes mellitus, hyperlipidemia, stroke, peripheral vascular disease, previous MI, CHF, smoking status, anterior infarct location, and aspirin, clopidogrel, β-blocker, and statin received within 24 hours before arrival. Before matching, mean propensity scores for No-OHS and OHS patients were 0.11326 and 0.07681 (standardized difference, 85%), respectively. The c-statistic of the propensity model was 0.69 and the events per variable (EPV) statistic exceeded 10. After matching, mean propensity scores were respectively 0.11341 and 0.11339 (standardized difference, 0.05%).
There were 4598 matched pairs. The propensity score matching was effective in achieving balance for each covariate used in the model. The absolute standardized differences between the 2 groups were <10% for all and <5% for most covariates (Table 1⇑). Only a history of stroke and prearrival treatment with clopidogrel were 5% or more. Patients presenting to OHS hospitals were significantly more likely to receive aspirin (91.7% versus 85.1%, P<0.001), clopidogrel (64.2% versus 53.8%, P=0.016), glycoprotein (Gp) IIb/IIIa inhibitors (60.7% versus 50.0%, P=0.036), β-blockers (82.6% versus 76.9%, P=0.006), and statins (51.1% versus 39.0%, P<0.001) within 24 hours after arrival but were less likely to receive a heparin product during their stay (78.7% versus 83.0%, P=0.030; Table 2).
Among reperfusion-eligible patients, those presenting to OHS hospitals were more likely to undergo pPCI (69.2% versus 44.5%, P<0.001) or any form of initial reperfusion therapy (81.5% versus 71.0%, P<0.001). Median but not mean door-to-balloon time was significantly shorter at OHS hospitals, and door-to-needle remained significantly longer than at No-OHS hospitals. The use of rescue PCI was similar (Table 3).
Among nontransfer patients, patients presenting to OHS hospitals had significantly lower mortality (7.2% versus 9.3%, P=0.025; Table 4). However, after adjusting for differences in medications administered within 24 hours of arrival, hospital characteristics, and the use of reperfusion therapy, the difference in mortality was of borderline significance (hazard ratio, 0.87; 95% CI, 0.75 to 1.01; P=0.063; Table 6).
At discharge, patients presenting to OHS hospitals were significantly more likely to be prescribed aspirin (93.1% versus 90.0%, P=0.008) and statins (82.9% versus 77.1%, P=0.016), and there was a trend toward increased use of β-blockers (90.7% versus 88.0%, P=0.055; Table 5).
Propensity-Matched Analysis, PCI Patients
The propensity matching was then restricted to patients who underwent primary or facilitated PCI (n=1,795 in each group). Before matching, mean propensity scores for No-OHS and OHS patients were 0.07274 and 0.06071 (standardized difference, 54%), respectively. The c-statistic of the propensity model was 0.62, and the EPV statistic exceeded 10. After matching, mean propensity scores were respectively 0.07301 and 0.07300 (standardized difference, 0.02%). Baseline characteristics, medication use, and presenting characteristics were well matched between the groups (Table 1⇑). The absolute standardized differences between the 2 groups were <10% for all and <5% for most covariates (Table 1⇑). Only a history of CHF and prearrival treatment with a statin were 5% or more (Table 1⇑).
In this propensity-matched analysis, there was no significant difference in the administration of guideline recommended medications between the groups (Table 2). Median but not mean door-to-balloon time was slightly shorter at OHS hospitals (Table 3). There was no significant difference in mortality (3.8% versus 3.3%, P=0.31), the composite of death and MI (4.5% versus 4.1%, P=0.72), or the composite of death, MI, CHF, and cardiogenic shock (16.5% versus 16.8%, P=0.56) between the groups (Table 4). Discharge medications were similar between the groups (Table 5).
STEMI patients presenting to interventional hospitals without on-site cardiac surgery have substantially higher in-hospital mortality, even after adjusting for differences in patient characteristics in a propensity analysis. This increase in mortality was of borderline significance after adjusting for differences in hospital and treatment characteristics, suggesting that efforts to increase adherence to guideline recommendations and use of immediate reperfusion therapy could close the gap in outcomes. Moreover, among patients undergoing pPCI, there was no significant difference in mortality.
A recent report from the National Cardiovascular Data Registry reported that in-hospital mortality was no different among 33 002 patients undergoing pPCI at OHS and No-OHS hospitals after adjustment for site characteristics and potential confounding variables.16 The results presented here corroborate the findings in the National Cardiovascular Data Registry report and suggest that pPCI is safe at hospitals without surgical back-up. However, in our propensity score model matching for baseline patient characteristics, more than a quarter of reperfusion-eligible patients presenting to No-OHS hospitals did not receive any form of reperfusion therapy, and fewer than half underwent primary or facilitated PCI. On the contrary, more than two thirds of patients presenting to OHS hospitals underwent primary or facilitated PCI, and fewer than 20% did not receive any form of reperfusion therapy.
On-site cardiac surgery may be a surrogate for quality of care, as patients presenting to No-OHS hospitals were significantly less likely to receive guideline-recommended medical therapies within 24 hours of arrival, even after adjusting for differences in baseline patient characteristics. In addition, larger hospital size and higher MI volume have been associated with improved outcomes among patients presenting with STEMI, and larger hospitals with higher MI volume are more likely to have back-up cardiac surgery.21,22
Early and sustained reperfusion of the infarct-related artery is the goal in STEMI.1–3 pPCI has been associated with improved clinical outcomes in randomized clinical trials.4–7 Clinical trials, however, often enroll a highly select group of patients who do not reflect what is frequently seen in real-world practice. In addition, these trials were all undertaken at hospitals with on-site cardiac surgery. As cardiac catheterization may have complications requiring urgent open-heart surgery, the safety of performing primary or facilitated PCI at hospitals without surgical back-up has been vigorously debated. Although the technology of PCI has rapidly advanced, evidence suggests there is no easy way to determine who will need urgent CABG after failed PCI.10
Other studies have demonstrated nonsignificant differences in clinical outcomes among STEMI patients presenting to OHS and No-OHS hospitals. In a study of Medicare data, Wennberg et al reported similar in-hospital mortality among patients undergoing pPCI or rescue PCI at No-OHS hospitals as at OHS hospitals after adjusting for differences in baseline clinical characteristics.14 In a Swedish registry that enrolled 4595 STEMI patients at OHS hospitals and 857 patients at No-OHS hospitals, 30-day mortality was 6.7% and 7.0%, respectively (P=NS).23
On the contrary, in a comparison of a single No-OHS hospital with a single OHS hospital from the Mayo clinic, STEMI patients presenting to the No-OHS hospital (n=285) had a significantly higher in-hospital mortality than those presenting to the OHS hospital (4% versus 1%, P=0.05).
As patients undergoing primary and facilitated PCI at No-OHS hospitals had a similar incidence of mortality as those at OHS hospitals, efforts to increase the use of pPCI at No-OHS may be warranted. This may be accomplished by making pPCI the default standard at all hospitals with PCI capability or providing 24-hour coverage of cardiac catheterization laboratories at all hospitals with PCI capabilities.
A major limitation of our study is the disparity in the number of transfer-out patients between OHS and No-OHS hospitals. However, patients transferred in to OHS hospitals were not substantially different from patients transferred from No-OHS hospitals in terms of baseline characteristics. The outcomes of transfer-out patients are not known and could alter the findings of the current study considerably. These patients may have a survival bias in that they were stable enough to be transferred. We attempted to control for this phenomenon in the propensity matching by including transfer-in status in the generation of the propensity score. That mortality remained significantly higher at No-OHS hospitals despite the similar number of transfer-in patients in the propensity-matched analysis suggests that processes independent of transfer-in status are the cause of the difference in mortality.
There are other limitations to the current study. Although many hospitals participated in NRMI, these hospitals may not be representative of all healthcare facilities, and they likely reflect larger more procedure-oriented centers. In addition, there was no independent validation of data forms. Finally, data regarding whether patients had contraindications to guideline-recommended medications was not collected in the NRMI data set.
STEMI patients presenting to hospitals without on-site cardiac surgery have substantially higher in-hospital mortality and are less likely to receive guideline-recommended medications within 24 hours and at discharge. The difference in mortality was of borderline significance after adjusting for hospital and treatment characteristics, suggesting the need for improved systems of care to increase adherence to guideline recommendations. Moreover, patients undergoing pPCI have similar mortality at hospitals with and without on-site cardiac surgery, suggesting that the difference in overall mortality is not attributable to complications of pPCI at hospitals without surgical back-up.
We appreciate the guidance of David P. Miller in the design of the statistical analysis.
Sources of Funding
The National Registry of Myocardial Infarction is supported by Genentech Inc, San Francisco, Calif.
Guest Editor for this article was Véronique L. Roger, MD, MPH.
Antman EM, Anbe DT, Armstrong PW, Bates ER, Green LA, Hand M, Hochman JS, Krumholz HM, Kushner FG, Lamas GA, Mullany CJ, Ornato JP, Pearle DL, Sloan MA, Smith SC Jr, Alpert JS, Anderson JL, Faxon DP, Fuster V, Gibbons RJ, Gregoratos G, Halperin JL, Hiratzka LF, Hunt SA, Jacobs AK. ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Revise the 1999 Guidelines for the Management of Patients with Acute Myocardial Infarction). Circulation. 2004; 110: e82–e292.
Gibson CM, Cannon CP, Murphy SA, Marble SJ, Barron HV, Braunwald E. Relationship of the TIMI myocardial perfusion grades, flow grades, frame count, and percutaneous coronary intervention to long-term outcomes after thrombolytic administration in acute myocardial infarction. Circulation. 2002; 105: 1909–1913.
Weaver WD, Simes RJ, Betriu A, Grines CL, Zijlstra F, Garcia E, Grinfeld L, Gibbons RJ, Ribeiro EE, DeWood MA, Ribichini F. Comparison of primary coronary angioplasty and intravenous thrombolytic therapy for acute myocardial infarction: a quantitative review. JAMA. 1997; 278: 2093–2098.
Nunn CM, O'Neill WW, Rothbaum D, Stone GW, O'Keefe J, Overlie P, Donohue B, Grines L, Browne KF, Vlietstra RE, Catlin T, Grines CL. Long-term outcome after primary angioplasty: report from the primary angioplasty in myocardial infarction (PAMI-I) trial. J Am Coll Cardiol. 1999; 33: 640–646.
Gibson CM, Pride YB, Frederick PD, Pollack CV Jr, Canto JG, Tiefenbrunn AJ, Weaver WD, Lambrew CT, French WJ, Peterson ED, Rogers WJ. Trends in reperfusion strategies, door-to-needle and door-to-balloon times, and in-hospital mortality among patients with ST-segment elevation myocardial infarction enrolled in the National Registry of Myocardial Infarction from 1990 to 2006. Am Heart J. 2008; 156: 1035–1044.
Weaver WD, Litwin PE, Martin JS. Use of direct angioplasty for treatment of patients with acute myocardial infarction in hospitals with and without on-site cardiac surgery. The Myocardial Infarction, Triage, and Intervention Project Investigators. Circulation. 1993; 88: 2067–2075.
Kutcher MA, Klein LW, Ou FS, Wharton TP Jr, Dehmer GJ, Singh M, Anderson HV, Rumsfeld JS, Weintraub WS, Shaw RE, SAcrinty MT, Woodward A, Peterson ED, Brindis RG. Percutaneous coronary interventions in facilities without cardiac surgery onsite: a report from the National Cardiovascular Data Registry. J Am Coll Cardiol. 2009; 54: 16–24.
Rogers WJ, Bowlby LJ, Chandra NC, French WJ, Gore JM, Lambrew CT, Rubison RM, Tiefenbrunn AJ, Weaver WD. Treatment of myocardial infarction in the United States (1990 to 1993). Observations from the National Registry of Myocardial Infarction. Circulation. 1994; 90: 2103–2114.
Rogers WJ, Canto JG, Lambrew CT, Tiefenbrunn AJ, Kinkaid B, Shoultz DA, Frederick PD, Every N. 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.
Thune JJ, Hoefsten DE, Lindholm MG, Mortensen LS, Andersen HR, Nielsen TT, Kober L, Kelbaek H. Simple risk stratification at admission to identify patients with reduced mortality from primary angioplasty. Circulation. 2005; 112: 2017–2021.
Wiviott SD, Morrow DA, Frederick PD, Giugliano RP, Gibson CM, McCabe CH, Cannon CP, Antman EM, Braunwald E. Performance of the thrombolysis in myocardial infarction risk index in the National Registry of Myocardial Infarction-3 and -4: a simple index that predicts mortality in ST-segment elevation myocardial infarction. J Am Coll Cardiol. 2004; 44: 783–789.
Canto JG, Every NR, Magid DJ, Rogers WJ, Malmgren JA, Frederick PD, French WJ, Tiefenbrunn AJ, Misra VK, Kiefe CI, Barron HV. The volume of primary angioplasty procedures and survival after acute myocardial infarction. National Registry of Myocardial Infarction 2 Investigators. N Engl J Med. 2000; 342: 1573–1580.
Carlsson J, James SN, Stahle E, Hofer S, Lagerqvist B. Outcome of percutaneous coronary intervention in hospitals with and without on-site cardiac surgery standby. Heart. 2007; 93: 335–338.