Six-Month Outcome in Patients With Myocardial Infarction Initially Admitted to Tertiary and Nontertiary Hospitals

Transcription

1 JACC Vol. 30, No. 5 November 1, 1997: Six-Month Outcome in Patients With Myocardial Infarction Initially Admitted to Tertiary and Nontertiary Hositals JAUME MARRUGAT, MD, GINÉS SANZ, MD,* RAFEL MASIÁ, MD, VICENTE VALLE, MD, LLUIS MOLINA, MD, MARIA CARDONA, MD,* JOAN SALA, MD, LLUIS SERÉS, MD, LLUIS SZESCIELINSKI, MD, XAVIER ALBERT, MD, JOSEP LUPÓN, MD, JORDI ALONSO, MD, FOR THE RESCATE INVESTIGATORS Barcelona and Girona, Sain Objectives. The aim of the resent study was to ascertain whether the degree of accessibility to coronary angiograhy and revascularization results in differing usages or outcomes, or both, in the setting of a high coverage national health system. Background. The selective use of coronary angiograhy and revascularization rocedures in the management of acute myocardial infarction (MI) remains controversial. Methods. A cohort of 1,460 consecutive atients with a first MI admitted to four referral teaching hositals (one with tertiary facilities) were followed u for 6 months after admission. Only atients initially admitted to each of the study hositals were retained for analysis in the original hosital s cohort. End oints were 6-month mortality and readmission for reinfarction, unstable angina, heart failure or severe ventricular arrhythmia. Results. Patients admitted to the tertiary hosital were more likely to undergo coronary angiograhy (adjusted relative risk 4.22, 95% confidence interval [CI] 3.37 to 5.45) than those admitted to the nontertiary sites (use rate: 22.1% for nontertiary care, 55.5% for tertiary care). Revascularization rocedures were erformed in 21.2% of atients in the tertiary hosital and in 8.3% in the nontertiary hositals ( < ). Median delay for emergency coronary angiograhy was shorter in the tertiary hosital (within 1 vs. 2 days, < ). Six-month mortality or readmission rates were similar (23.7% and 24.7% for tertiary and nontertiary care, resectively). After adjustment for comorbidity and disease severity, the relative risk of death or readmission for the tertiary hosital was 1.03 (95% CI 0.69 to 1.53) times that of the nontertiary hositals. Conclusions. Selective use of coronary angiograhy and revascularization rocedures may be as effective as less restricted use in the management of acute MI. (J Am Coll Cardiol 1997;30: ) 1997 by the American College of Cardiology The rognosis of atients admitted for acute myocardial infarction (MI) has rogressively imroved in the ast 30 years, due mainly to the develoment of coronary care units (1) and the use of thrombolytic theray (2,3) and other harmacologic treatments (4,5). However, the otimal use rate of tertiary care rocedures, such as coronary angiograhy, coronary artery byass graft surgery (CABG) (6) and ercutaneous transluminal coronary angiolasty (PTCA) (7), has not been established. Marked geograhic variations in use rate have been described From the Deartment of Eidemiology and Public Health, Institut Municial d Investigació Mèdica, Barcelona; *Institute of Cardiovascular Diseases, Hosital Clinic, Barcelona; Deartment of Cardiology, Hosital Jose Trueta, Girona; Deartment of Cardiology, Hosital Germans Trias, Badalona; and Deartment of Cardiology, Hosital del Mar, Barcelona, Sain. A comlete list of the Recursos Emleados en el Síndrome Coronario Agudo y Tiemos de Esera (RESCATE) Investigators aears in the Aendix. This roject was funded by Grant 92/0009 from the Fondo de Investigación Sanitaria, Madrid and by Grant CIRIT/SGR from the Generalitat de Catalunya, Barcelona, Sain. Manuscrit received Aril 23, 1997; revised manuscrit received July 17, 1997, acceted July 21, Address for corresondence: Dr. Jaume Marrugat, Deartament d Eidemiologia i Salut Pública, Institut Municial d Investigació Mèdica (IMIM), Carrer Doctor Aiguader 80, E Barcelona, Sain. jaume@imim.es. (8,9), and on-site availability is one of the strongest redictors of their use (10). It remains to be determined whether restricted use of such rocedures results in worse atient outcome. Retrosective studies (11,12) suggest that a high use rate of tertiary rocedures does not result in better survival, although anginal symtoms may be reduced and quality of life and functional status at 1 or 2 years imroved. The Sanish National Health System covers 97% of the oulation (13); thus, hosital treatment does not deend on the atient s ability to ay. The aims of the resent study were to ascertain whether the variation in accessibility to coronary angiograhy, PTCA and CABG, deending on the in-hosital availability of these rocedures, determines different use rates or delays in atients with MI and to assess whether these differences, if existent, are associated with differences in outcome. To discuss this article on-line, visit the ACC Home Page at and click on the JACC Forum 1997 by the American College of Cardiology /97/$17.00 Published by Elsevier Science Inc. PII S (97)

2 1188 MARRUGAT ET AL. JACC Vol. 30, No. 5 SIX-MONTH OUTCOME AFTER FIRST MI November 1, 1997: Abbreviations and Acronyms CABG coronary artery byass graft surgery CI confidence interval MI myocardial infarction PTCA ercutaneous transluminal coronary angiolasty Methods Study design. The study was designed as a 6-month follow-u study of atients admitted to one hosital with and three without angiograhy or coronary surgery facilities. All four articiating hositals were ublic teaching institutions. Patients admitted to the tertiary care hosital were referred to as grou A and those admitted to the nontertiary hositals as grou B. The tertiary hosital included exclusively rimarily admitted atients; thus, atients referred from grou B hositals were not included. Patients from grou B hositals were referred to several tertiary hositals in Barcelona for angiograhic rocedures, and each atient outcome was attributed to the initial admitting hosital. Inclusion criteria. Between May 1992 and June 1994, all atients with a first MI u to the age of 80 years admitted to the four articiating hositals within 72 h of onset of symtoms of MI were included. MI was diagnosed when two of the following criteria were resent: 1) abnormal new Q waves, 2) increase in cardiac enzyme levels (more than twice the uer normal value), and 3) tyical chest ain 20 min in duration. Exclusion criteria. Residence outside the study areas or any of the following conditions: 1) life-threatening diseases other than the index event; 2) revious CABG or PTCA; 3) or coronary angiograhy in the ast 6 months. Patients enrolled in ongoing clinical trials were not excluded to reroduce actual care scenarios more faithfully. Primary end oints. The comosite rimary end oint included mortality or readmission within 6 months of the onset of MI for any of the following reasons: 1) reinfarction, 2) congestive heart failure, 3) ventricular fibrillation or tachycardia, or 4) unstable angina. Reinfarction was defined as a new infarction occurring at least 28 days after the onset of the initial event. Congestive heart failure was diagnosed clinically according to standard diagnostic criteria (14,15). Ventricular tachycardia was considered an end oint only when sustained and leading to hosital admission. Progressive and rest angina were considered unstable angina, therefore requiring hosital admission according to Braunwald criteria (16). Samle size. Samle size was chosen to obtain a statistical ower of 0.80 in a two-tailed test with an alha risk of 0.05 if a difference 10 ercentage oints in the 6-month event rate was observed between the tertiary and the nontertiary hositals (20% and 30% of rimary end oints, resectively). A 10% increase in the intended samle was alied to comensate for atients lost to follow-u; thus, 1,300 atients were required, of whom at least 325 had to be admitted to Hosital A. This samle size would ermit a relative risk 1.5 to be statistically significant ( 0.05). Management of MI. Each articiating hosital was allowed to follow its own routine, and no attemt was made to standardize atient management. However, all four hositals had written MI rotocols in accordance with international guidelines (17 19). Aroriateness of rocedures. Emergency indications for coronary angiograhy, PTCA and CABG were standardized in advance to assess the need for their urgent use. These criteria were adated from secific international treatment guidelines (17 19). Emergency coronary angiograhy was considered aroriate in the resence of 1) recurrent eisodes of angina, articularly if accomanied by ST-T wave changes, not controlled after 48 h of aroriate treatment; or 2) mechanical comlications, including severe mitral regurgitation due to aillary muscle dysfunction or ventricular setal ruture. Emergency revascularization was considered aroriate in either of the aforementioned circumstances when coronary anatomy was deemed suitable. CABG was referred to PTCA in atients with left main coronary artery stenosis or diffuse coronary disease (two to three vessels) or when cardiac reair was necessary. Patients with one- or two-vessel discrete lesions were judged to be candidates for PTCA. The need for elective angiograhy was assessed in detail in a random subsamle half the total samle size. This rocedure was considered necessary if at least one of the following occurred: 1) ostinfarction angina, 2) mechanical comlications, 3) ositive exercise test results, or 4) reinfarction within 28 days. Study variables in acute hase of MI. The following variables were rosectively recorded by a trained investigator at each center: demograhic data; history of hyertension; diabetes; chronic obstructive ulmonary disease; eriheral vascular disease; smoking status; MI location; Killi class; resence of severe arrhythmia (defined as the occurrence of at least one eisode of ventricular fibrillation or sustained ventricular tachycardia requiring immediate medical intervention) within the first 72 h; delay from onset of symtoms to first monitoring in an emergency room, coronary care unit or general intensive care unit; and hosital stay, use of thrombolysis, exercise test, coronary angiograhy, PTCA and CABG and the comlications associated with diagnostic and theraeutic rocedures. Analysis and statistical methods. Grous A and B were assessed for differences in categoric variables by the chi-square or Fisher exact test when aroriate and by the Student t or Mann-Whitney U test when necessary for differences in continuous variables. The level of significance used was 5%. Survival curves were estimated by the Kalan-Meier method. Adjusted relative risks for 6-month mortality and morbidity were estimated using unconditional logistic regression (20). Severity or rognosis-related variables showing interhosital differences were adjusted for in the models to control for case mix. All two-level interactions between airs of these variables were assessed in all models. The SPSS and

3 JACC Vol. 30, No. 5 November 1, 1997: MARRUGAT ET AL. SIX-MONTH OUTCOME AFTER FIRST MI 1189 EGRET (Statistics and Eidemiology Research Cororation) statistical ackages were used. Results Patients. Of the 2,397 atients registered in the four hositals, 1,460 (60.8%) (1,035 initially admitted to nontertiary hositals, 425 to tertiary hosital) met the inclusion criteria. Previous MI (18.4%) and age 80 years (9.6%) were the most frequent reasons for exclusion. Other causes included atients residing outside the hosital catchment area, which made follow-u imractical (4.5%); revious revascularization (1.0%); and angiograhy in the revious 6 months (0.3%). An additional 5.8% were excluded for miscellaneous reasons, such as administrative; terminal, severe noncoronary disease; and referral from articiant hositals. Patients admitted to the three grou B hositals did not differ clinically or statistically in any relevant characteristic (results not shown). Differences in some demograhic and clinical variables were found between grous A and B (Table 1). The grou A hosital (the only tertiary site) more frequently admitted atients with comorbidity (i.e., diabetes and hyertension), non-q wave MI or revious angina than did the grou B hositals. Conversely, atients in grou B were more often in Killi class III or IV than atients in grou A. Procedures. Median delays to first cardiac monitoring were similar in grous A and B. More than 50% of atients erformed an exercise test. Thrombolytic theray was administered in 545 atients (37.4%) (Table 2), with a median time from symtom onset to administration of 3 h. Stretokinase was the drug of choice in 77.8% of occasions. Coronary angiograhy was erformed in 463 atients (31.8%), PTCA in 93 (6.5%) and CABG in 83 (5.7%) within 6 months of admission. By the end of the follow-u eriod, the Table 1. Demograhic and Clinical Characteristics of Study Patients Admitted to Tertiary (grou A) and Nontertiary Hositals (grou B) Grou A Grou B Age (yr) NS Women NS Killi class III or IV Anterior MI NS Non-Q wave MI Arrhythmia* NS COPD NS Diabetes Hyertension PVD NS Smoker NS Previous angina *Ventricular fibrillation within 72 h of onset of symtoms of myocardial infarction (MI). At least 1 cigarette/day. Data resented are mean value SD or ercent of atients. COPD chronic obstructive ulmonary disease; PVD eriheral vascular disease. Table 2. Time to Coronary Care Unit Admission, Length of Hosital Stay and Use of Coronary Angiograhy and Revascularization Procedures by Hosital Tye Grou A (tertiary hosital) Grou B (nontertiary hositals) Time from symtom onset to monitoring (h) Median NS Range 0 to 24 0 to 24 Time from symtom onset to CCU admission (h) Median Range 0.8 to to 72 CCU stay (days) Median Range 1 to 37 1 to 60 Hosital stay (days) Median Range 1 to to 96 Thrombolysis NS Exercise test NS 6-mo follow-u Elective coronary angio PTCA CABG Data resented are ercent of atients, unless otherwise indicated. angio angiograhy; CABG coronary artery byass graft surgery; CCU coronary care unit; PTCA ercutaneous transluminal coronary angiolasty. tertiary hosital had erformed more coronary angiograhy than the nontertiary hositals (55.5% vs. 22.1%, ), and the use rate of elective angiograhy was higher in the tertiary hosital (50.1% vs. 14.9%, ) (Table 2). In the random subsamle, these rocedures were deemed necessary in 40.7% of atients in the tertiary hosital and 39.6% in the nontertiary hositals ( NS). The roortion of atients who underwent revascularization was higher in grou A than in grou B (PTCA: 12.0% vs. 4.1%, resectively, ; CABG: 9.2% vs. 4.3%, resectively, ; PTCA or CABG: 21.0% vs. 8.3%, resectively, ) (Table 2). In a model adjusted for age, Killi class, gender, diabetes, hyertension, MI location, Q wave MI and revious angina, the relative risk for coronary angiograhy use among atients in grou A versus grou B was 4.22 (95% confidence interval [CI] 3.27 to 5.45). Use rate of angiograhy in atients with a non-q wave MI was 1.34 (95% CI 0.97 to 1.99) times that of atients with a Q wave MI. Comlications. Comlications occurring during the 6- month eriod were death (one atient [0.3%]), MI (four atients), stroke (two atients) and surgical femoral artery reair (six atients) for coronary angiograhy; death (one atient) (1.4%) and surgical reair (one atient) for PTCA; and death (three atients [4.7%]), MI (one atient) and stroke (one atient) for CABG. Aroriateness and delay in urgent rocedure use. Rates of emergency coronary angiograhy and revascularization

4 1190 MARRUGAT ET AL. JACC Vol. 30, No. 5 SIX-MONTH OUTCOME AFTER FIRST MI November 1, 1997: Table 3. Emergency Coronary Angiograhy and Revascularization Within 28 Days of Onset of Myocardial Infarction in Patients Meeting Standardized Criteria for These Procedures in Tertiary (grou A) and Nontertiary Hositals (grou B) Grou A Grou B Urgent angiograhy Pts meeting criteria 22 (5.2%) 74 (7.2%) NS Performed NS* Delay Median Range Urgent PTCA Pts meeting criteria 3 (0.7%) 11 (1.1%) NS Performed 3 10 NS* Delay Median 0 0 NS Range Urgent CABG Pts meeting criteria 17 (4.0%) 26 (2.5%) NS Performed NS* Delay Median 1 0 NS Range *Fisher exact test. Mann-Whitney U test. Data resented are number (%) of atients, unless otherwise indicated. Abbreviations as in Table 2. techniques according to hosital tye are shown in Table 3. Ninety-six atients (6.4%) met objective criteria for emergency coronary angiograhy, which was eventually erformed in 77 (80.8%). No differences between grous were observed in the roortion of atients in whom PTCA was indicated or erformed. In 19 atients, emergency coronary angiograhy was indicated but not erformed: In 2 of these 19 it was not requested by the attending hysician (both atients were alive at the end of the 6-month follow-u eriod). Of the remaining 17 atients, 12 raidly deteriorated and died before the rocedure could be erformed. In five atients referred by nontertiary hositals, coronary angiograhy was not erformed by the receiving tertiary hosital (four atients were alive at end of the 6-month follow-u). Among the 76 atients with emergency indication for coronary angiograhy, the roortion of atients meeting emergency criteria for PTCA or CABG and the roortion of atients who finally underwent these rocedures was similar in both grous. The median delay in erforming emergency catheterization was shorter in the tertiary hosital than in the nontertiary hositals (within 1 vs. 2 days, resectively, 0.001) (Table 3). Outcome. Both 28-day and 6-month mortality and readmission rates for both grous are shown in Table 4. Only two atients were lost to follow-u. Overall intergrou differences in mortality or readmission rate were not statistically significant. Survival curves in both grous were similar and not statistically significant (Fig. 1). Logistic regression models adjusted for differences between Table 4. Primary End Points During Follow-U in Study Patients Admitted to Tertiary (grou A) or Nontertiary Hositals (grou B) Grou A Grou B Death Overall 6 mo 60 (14.2) 147 (14.2) NS 28 day 44 (10.4) 110 (10.6) NS 6 mo* 16 (4.2) 37 (4.0) NS Readm* Reinfarction 11 (2.9) 35 (3.9) NS Angina 25 (6.6) 64 (7.0) NS VF/VT 1 (0.3) 15 (1.6) Cardiac failure 12 (3.2) 40 (4.4) NS Any of above 43 (11.7) 127 (14.0) NS Death or Readm 99 (23.7) 252 (24.7) NS *Among 28-day survivors. Data resented are number (%) of atients. Readm hosital readmission; VF/VT ventricular tachycardia/ventricular fibrillation. grous A and B (Table 5) show that hosital tye was not an indeendent risk factor for 6-month mortality or readmission. The relative risk in grou B was 0.97 (95% CI 0.68 to 1.55). No statistically significant interaction terms were identified. No differences between grous were found when only 6-month mortality was considered as a deendent variable. The relative risk for grou B was 1.12 (95% CI 0.83 to 1.51). No statistically significant interaction terms were identified (Table 5). Discussion The results of the resent study show that the use of tertiary rocedures in atients with a first acute MI varies according to the tye of admitting hosital. Although no differences were found in the roortion of atients undergoing emergency diagnostic or theraeutic rocedures, longer delays were encountered in atients admitted to nontertiary hositals. Furthermore, admission to an institution with on-site cardiac Figure 1. Survival robability for tertiary (dashed curve) versus nontertiary hositals (solid curve) ( 0.98).

5 JACC Vol. 30, No. 5 November 1, 1997: MARRUGAT ET AL. SIX-MONTH OUTCOME AFTER FIRST MI 1191 Table 5. Adjusted Relative Risks and 95% Confidence Intervals for 6-Month Mortality After First Myocardial Infarction in Patients Admitted to Three Hositals Without On-Site In-Hosital Catheterization Laboratory and Coronary Surgery 6-mo Mortality [RR (95% CI)] 6-mo Mortality/Readm [RR (95% CI)] Nontertiary hositals* 0.97 ( ) 1.06 ( ) Killi class III or IV ( ) 7.91 ( ) Diabetes 1.59 ( ) 1.31 ( ) Hyertension 1.25 ( ) 1.29 ( ) Previous angina 1.39 ( ) 1.48 ( ) Non-Q wave MI 0.56 ( ) 0.54 ( ) *Reference grou: tertiary hosital. CI confidence interval; RR relative risk; other abbreviations as in Tables 1 and 4. catheterization facilities was strongly associated with the use of coronary angiograhy. However, no differences were found in the 6-month mortality or readmission rate between the tertiary and the three nontertiary hositals, even after adjustment for severity and comorbidity variables. These results raise the question of whether differences in use reflect an excess in the number of rocedures erformed in the tertiary hosital. Several attemts have been made to assess the consequences of differences in acute coronary syndrome management, with inconclusive results (21 23). Large geograhic variations in rocedure use rate have also been described, which indicates that some use may be inaroriate (8,24,25). One study suggests that 30% of CABG may be erformed for equivocal reasons and that 14% are inaroriate (26) and another that low risk atients are selected for angiograhy and revascularization (27). Interestingly, in our study the roortion of atients with objective indications for emergency coronary angiograhy, PTCA and CABG was similar in both hosital tyes, although longer delays were observed in the nontertiaries. Therefore, the differences observed in utilization rates were due to the larger number of elective rocedures carried out in the tertiary hosital. Again, the roortion of elective angiograms deemed necessary was similar in both tyes of hositals when it was investigated in a random subsamle. Patients admitted to hositals with on-site cardiac catheterization facilities are more likely to undergo coronary angiograhy than atients admitted to other hositals (10). The results of our study concur with that finding, also confirmed by others (27,28). The greater number of atients with non-q wave MI admitted to the tertiary hosital may in art account for the higher use of elective angiograhy. In addition to availability, financial incentives and atient demand have been suggested as factors influencing variations in the use of tertiary rocedures in atients with coronary disease (11). Hositals in the Sanish National Health System do not rovide financial incentives for hysicians to erform rocedures. It is therefore highly unlikely that financial reasons accounted for the differences observed in our study. Study characteristics and limitations. The finding that outcome was similar in both hosital tyes and was unrelated to the use of invasive rocedures is in accordance with reviously reorted results (11). The resent study was secifically designed and owered to detect differences in event rates between tertiary and nontertiary hositals. The observation eriod was extended to 6 months, when most of the events related to the acute hase of MI would already have occurred. However, according to the results of revious studies (11,12), a longer observation eriod might reveal differences in anginal symtoms or functional status. The resent study did not address other issues related to on-site unavailability of tertiary care rocedures, such as inconvenience and distress for atients transferred between hositals and their relatives. The following measures were taken to ensure efficient case mix control: Only atients with a first MI were included; exclusions and their causes had to be justified; and statistical adjustment for case mix (i.e., differences in disease severity and comorbidity between the two hosital tyes) was used. To revent hysician-deendent outcomes, PTCA or CABG after discharge was not used as an end oint. Conclusions. Our results suggest that desite longer delays and robably more inconvenience to atients, the selective use of coronary angiograhy and revascularization rocedures alied in nontertiary centers is as effective as the less restricted use observed in tertiary hositals. We areciate the English revision of the manuscrit made by Christine O Hara. Aendix Particiating Institutions and Investigators for the Recursos Emleados en el Síndrome Coronario Agudo y Tiemos de Esera (RESCATE) Study G. Sanz, M. Cardona, Hosital Clinic i Provincial de Barcelona; R. Masiá, J. Sala, X. Albert, Hosital Jose Trueta de Girona; L. Molina, L. Szescielinski, Hosital del Mar de Barcelona; V. Valle, A. Curós, L. Serés, J. Luón, J. Serra, D. Pereferrer, Hosital Germans Trias i Pujol de Badalona; J. Marrugat, J. Alonso, J. Vila, M. Pavesi, Institut Municial d Investigació Mèdica de Barcelona. References 1. Brown KWG, MacMillan RL, Forbath N, et al. Coronary unit: an intensive care centre for acute myocardial infarction. Lancet 1963;2: Gruo Italiano er lo Studio della Stretochinasi nell Infarto Miocardico (GISSI). Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Lancet 1986;1: GUSTO Investigators. An international randomized trial comaring four thrombolytic strategies for acute myocardial infarction. N Engl J Med 1993;329: The Beta Blocker Pooling Project Research Grou. The Beta Blocker Pooling Project (BBPP): subgrou findings from randomized trials in ostinfarction atients. Eur Heart J 1988;9: Asirin Myocardial Infarction Study Research Grou. A randomized, controlled trial of asirin in ersons recovered from a myocardial infarction. JAMA 1980;243: Willerson JT, Frazier OH. Reducing mortality in atients with extensive myocardial infarction. N Engl J Med 1991;325:

6 1192 MARRUGAT ET AL. JACC Vol. 30, No. 5 SIX-MONTH OUTCOME AFTER FIRST MI November 1, 1997: Tool E. The thrombolysis and angiolasty in acute myocardial infarction (TAMI) trial. In: Acute Coronary Intervention. Tool EJ, editor. New York: Alan R. Liss, 1988: Chassin MR, Kosecoff J, Park RE, et al. Does inaroriate use exlain geograhic variations in the use of health care services? A study of three rocedures. JAMA 1987;258: Brook RH, Kosecoff JB, Park RE, Chassin MR, Winslow CM, Hamton JR. Diagnosis and treatment of coronary disease: comarison of doctors attitudes in the USA and the UK. The Lancet 1988;2: Every NR, Larson EB, Litwin PE, et al. The association between on-site cardiac catheterization facilities and the use of coronary angiograhy after acute myocardial infarction. N Engl J Med 1993;329: Rouleau JL, Moye LA, Pfeffer MA, et al., and the SAVE Investigators. A comarison of management atterns after acute myocardial infarction in Canada and the United States. N Engl J Med 1993;328: Mark DB, Naylor CD, Hlatky MA, et al. Use of medical resources and quality of life after acute myocardial infarction in Canada and the United States. N Engl J Med 1994;331: Borrell C, Pasarín I, Plasència A. Enquesta de salut de Barcelona, Barcelona: estadístiques de Salut 23, Ajuntament de Barcelona, 1995: McKee PA, Castelii WP, McNamara PM, Kannel WB. The natural history of congestive heart failure: the Framingham Heart Study. N Engl J Med 1971;285: Killi T, Kimball JT. Treatment of myocardial infarction in a coronary care unit; a two-year exerience with 250 atients. Am J Cardiol 1967;20: Braunwald E. Unstable angina: a classification. Circulation 1989;80: Bosch X, Fernández F, Bermejo J, Curós A, Valentín V. Tratamiento del infarto agudo de miocardio no comlicado. Rev Es Cardiol 1994;47 Sul: Froufe J, Lóez-Sendón J, Figueras J, Dominguez JM, Quintana JF, Vazquez C. Infarto agudo de miocardio comlicado. Rev Es Cardiol 1994;47 Sul: ACC/AHA guidelines for the early management of atients with acute myocardial infarction. Circulation 1990;82: Hosmer DW, Lemeshow S. Alied Logistic Regression. New York: John Wiley, DeFriese G. Measuring the effectiveness of medical interventions: new exectations of Health Services Research [editorial]. Health Serv Res 1990;25: Pashos C, McNeil B. Consequences of variation in treatment for acute myocardial infarction. Health Serv Res 1990;25: Lee TH, editor. Evaluating the Quality of Cardiovascular Care: A Primer. Bethesda (MD): American College of Cardiology, Pilote L, Califf RM, Sa S, et al., for the GUSTO Investigators. Regional variation across the United States in the management of acute myocardial infarction. N Engl J Med 1995;333: Leae LL, Park RE, Solomon DH, Chassin MR, Kosekoff J, Brook RH. Does inaroriate use exlain small-area variations in the use of health care services? JAMA 1990;263: Winslow CM, Kosecoff JB, Chassin M, Kanouse E, Brook RH. The aroriateness of erforming coronary artery byass surgery. JAMA 1988; 260: Pilote L, Miller DP, Califf RM, Rao JS, Weaver WD, Tool EJ. Determinants of the use of coronary angiograhy and revascularization after thrombolysis for acute myocardial infarction. N Engl J Med 1996;335: Selby JV, Fireman BH, Lundstrom RJ, et al. Variation among hositals in coronary angiograhy ractices and outcomes after myocardial infarction in a large health maintenance organization. N Engl J Med 1996;335: