Author + information
- Received December 3, 1998
- Revision received June 24, 1999
- Accepted August 27, 1999
- Published online December 1, 1999.
- Josep Lupón, MDa,
- Vicente Valle, MDa,
- Jaume Marrugat, MD∗,* (, )
- Roberto Elosua, MD∗,
- Lluis Serés, MDa,
- Marco Pavesi, PhD∗,
- Román Freixa, MDa,
- Ginés Sanz, MD†,
- Rafel Masiá, MD‡,
- Lluis Molina, MD¶,
- Joan Sala, MD‡,
- Jordi Serra, MDa,
- for the R.E.S.C.A.T.E. Investigators
- ↵*Reprint requests and correspondence: Dr. Jaume Marrugat, Departament d’Epidemiologia i Salut Pública, Institut Municipal d’Investigació Mèdica (IMIM), Carrer Doctor Aiguader 80, E-08003 Barcelona, Spain
The study assessed whether varying accessibility of patients with unstable angina (UA) to coronary angiography and revascularization determined differing usages and outcomes.
The appropriate use rate of coronary angiography and revascularization procedures in UA remains to be established.
A total of 791 consecutive patients with UA without previous acute myocardial infarction (AMI) admitted to four reference teaching hospitals (one with tertiary facilities) were followed for six months. End points were six-month mortality and readmission for AMI, UA, heart failure, or severe ventricular arrhythmias.
Patients admitted to the tertiary hospital were 3.27 (95% confidence interval [CI] 2.32 to 4.62) times more likely to undergo coronary angiography after adjustment for comorbidity and severity than were those admitted to nontertiary facilities (overall six-month use rates 70.1% and 48.3%, respectively). Revascularization procedures were performed in 36.2% of patients in the tertiary hospital and 24.6% in the others (p = 0.0007); adjusted relative risk (RR) 2.37 (95% CI 1.55 to 3.63). Median delay for urgent coronary angiography was shorter in the tertiary hospital (24 h vs. 4 days, p < 0.0002). Six-month mortality and readmission rates were similar in tertiary and nontertiary hospitals: 3.9% versus 5.3% and 16.9% versus 21.2%, respectively. Adjusted RR of death or readmission for the nontertiary hospitals was 1.23 (95% CI 0.57 to 2.67).
The use of coronary angiography and revascularization procedures in UA patients with no previous AMI is higher in tertiary than in nontertiary hospitals, but the more selective use of these procedures in nontertiary centers does not imply worse outcome.
The appropriate use rate of coronary angiography, coronary artery bypass grafting (CABG) and percutaneous transluminal coronary angioplasty (PTCA) in the management of acute coronary syndromes has not been definitively established (1). Their use in unstable angina (UA) is particularly controversial (2,3). On-site availability is one of the strongest predictors of tertiary procedure use rate in acute myocardial infarction (AMI) patients (4–6). Consequently, it would be of great interest to determine whether on-site availability also leads to a higher use rate of such procedures in UA patients, and whether this benefits patient prognosis.
The Spanish National Health System covers close to 100% of the population; hospital treatment therefore does not depend on the patient’s ability to pay, and direct economic stimulus to perform any particular procedure does not exist.
The aims of this study were to assess whether in-hospital availability of coronary angiography, PTCA and CABG determines different use rates or delays in UA patients and to assess whether these differences, if they exist, are associated with differences in outcome.
The RESCATE (Recursos Empleados en el Sı́ndrome Coronario Agudo y Tiempos de Espera) study consisted of a registry of first AMI and UA patients admitted to one hospital with, and three others without, coronary angiography facilities or coronary surgery. Patients were followed for six months after admission.
All four participating hospitals were teaching institutions. Patients admitted to the tertiary hospital are referred to as group A and those admitted to nontertiary as group B. The tertiary hospital included exclusively primarily admitted patients. Patients from group B hospitals were referred to several tertiary hospitals (including the study group A hospital) for angiographic and revascularization procedures. However, these procedures and each patient outcome were attributed to the initial admitting hospital.
Between May 1992 and June 1994, all primary UA patients up to the age of 80 years with no history of myocardial infarction admitted to the four participating hospitals were included.
The diagnosis of UA was made when typical chest pain occurred in any of the following presentations: 1) progressive angina (i.e., increase in the number of angina pectoris attacks or progressive decrease in physical exertion in the last month); 2) angina at rest (i.e., ischemic-type chest pain at rest of less than 20-min duration); 3) prolonged angina (i.e., ischemic-type chest pain lasting more than 20 min); and 4) variant angina (i.e., ischemic-type chest pain at rest with ST-segment elevation). Any one of these four types was considered to be new-onset angina when it lasted less than one month. However, new-onset angina per se was not considered unstable if it did not meet criteria for one category of the above classification. Conversely, ischemic electrocardiographic (ECG) changes during symptoms at any time of hospitalization, positive exercise test, significant lesions at coronary angiography or previous diagnosis of angina also had to be present. The diagnosis of AMI was ruled out in all patients by serial enzymatic determinations. Informed consent was obtained from all patients before their inclusion in the cohort, and the study was approved by the ethics committee of the four participating hospitals.
Exclusion criteria included previous AMI, residence outside the catchment areas, previous inclusion in the registry or any of the following conditions: life-threatening diseases other than the index event, previous CABG or PTCA, or coronary angiography in the last six months. Patients enroled in ongoing clinical trials were not excluded so as to reproduce more faithfully the real caring scenarios.
Primary end points
A composite primary end point included mortality or readmission within six months after the onset of UA for any of the following reasons: AMI, UA, congestive heart failure, sustained ventricular tachycardia or ventricular fibrillation.
Sample size was chosen to obtain a statistical power of 0.80 in a two-tailed test with an alpha risk of 0.05 if a difference greater than or equal to 10% units in the six-month event rate was observed between the tertiary hospital and the other facilities (15% and 25% of primary end points, respectively). A 10% increase in the intended sample was applied to compensate for patients lost to follow-up. In such conditions, 664 patients were required, at least 166 of whom were admitted to Hospital A. Arcsine rather than chi-square-derived formula was used to obtain a more conservative sample size calculation. This sample size would permit relative risks (RRs) greater than or equal to 1.7 to be statistically significant (p < 0.05).
Unstable angina management
All four hospitals followed similar medical management for UA, according to treatment widely used in clinical practice, including antiplatelet drugs, intravenous (IV) heparin, beta-blockers, IV and oral nitrates, and calcium antagonists, but no standard treatment of patients was established.
Appropriateness of procedures
Coronary angiographywas indicated as urgent in the presence of recurrent episodes of angina, particularly if accompanied by ST–T-wave changes not controlled after 48 h of appropriate treatment. In this setting, urgent revascularization was indicated when coronary anatomy was deemed suitable.
Elective angiographywas considered appropriate when, despite adequate medical control of symptoms, the exercise test was positive at less than 5 metabolic equivalents (METs) or when elevation of ST-segment was present during angina episodes. Revascularization, when undertaken in this setting, was considered elective.
In patients with left main coronary artery stenosis or diffuse coronary disease (2 to 3 vessels), CABG was preferred to PTCA. Patients with single or two-vessel discrete lesions were judged candidates for PTCA.
The following variables were prospectively recorded by a trained medical investigator at each center: demographic data, history of hypertension, diabetes, chronic obstructive pulmonary disease, peripheral vascular disease, smoking status, previous angina, acute pulmonary edema or cardiogenic shock, ECG changes during admission, presence of severe arrhythmia (defined as the occurrence of at least one episode of sustained ventricular tachycardia requiring immediate medical intervention or ventricular fibrillation), delay from onset of symptoms to first monitoring in an emergency room, hospital stay, exercise test, coronary angiography, PTCA and CABG.
Analysis and statistical methods
Differences between groups A and B were assessed for categorical variables by chi-square test or the Fisher exact test when appropriate, and by the Student ttest or Mann-Whitney Utest when necessary for continuous variables. The level of significance used was 5%.
Survival curves were estimated by the Kaplan-Meier method. Adjusted RRs for six-month mortality and morbidity were estimated using unconditional logistic regression (7). Severity or prognosis-related variables showing interhospital differences were adjusted in the models to control for case-mix. All two-level interactions between pairs of these variables were assessed in all models. The SPSS statistical package was used.
Of the 2,661 patients with UA admitted to the four hospitals, 839 were initially considered to meet the inclusion criteria. Previous AMI (34%) and previous inclusion (12%) were the most frequent reasons for exclusion. Other causes included age over 80 years (6%), previous revascularization or angiography in the previous six months (6%) and patients residing outside the hospital catchment areas that made follow-up impractical (2%). An additional 8% were excluded for miscellaneous reasons such as administrative motives, terminal severe noncoronary disease and referral from participant hospitals. A further 48 patients were excluded owing to insufficient evidence of ECG changes, positive exercise test, coronary lesions at angiography or previous definite angina diagnosis. Thus, 791 patients were finally retained for analysis.
Differences in some demographic and clinical variables were found between groups A and B. Hospital A (the only tertiary facility) admitted more frequently those patients with hypertension and previous angina than hospitals in the B group. Conversely, patients in group B more frequently developed acute pulmonary edema or cardiogenic shock and ECG changes during admission and had a greater number of angina crises in the 24 h before admission than did group A patients (Table 1).
Median delays in receiving first cardiac monitoring were similar in groups A and B. Total hospital stay was similar in both groups (Table 2). Overall, more than half the patients performed an exercise test, but a significantly higher proportion did so in the group B hospitals (Table 2).
A total of 448 coronary angiograms (56.6%), 133 PTCAs (16.8%) and 109 CABGs (13.8%) were performed within six months’ postadmission. At the end of follow-up, the tertiary hospital had performed more coronary angiograms than did the nontertiary hospitals (70.1% vs. 48.3%, p < 0.0001). In the model adjusted for the basal variables that differed between the two types of hospital (acute pulmonary edema/cardiogenic shock, hypertension, number of angina crises in the 24 h before admission, ECG changes and previous angina), the RR for coronary angiography use among patients in group A was 3.27 (95% confidence interval [CI] 2.32 to 4.62) compared with group B. The proportion of patients who received PTCA was higher in group A than in group B (22.0% vs. 13.6%, p = 0.002), but the proportion of patients receiving CABG was not statistically different (16.4% vs. 12.1%, p = 0.0855) (Table 2). The use of either revascularization method was 36.2% and 24.6%, respectively (p = 0.0007), adjusted risk 2.37 (95% CI 1.55 to 3.63). The use rate of elective coronary angiograms was higher in the tertiary hospital (95.3% vs. 70.6%, p < 0.0001) (Table 3). However, the use rate of urgent coronary angiography did not differ, although the proportion of patients meeting the defined urgent criteria was higher in the nontertiary hospitals (Table 3).
Appropriateness and delay in procedure use rate
Rates of urgent coronary angiography and revascularization techniques per type of hospital are shown in Table 3. Eighty-eight patients (11.1%) met protocol criteria for urgent coronary angiography, which was eventually performed in 79 patients (89.8%). No differences were observed between groups in the proportion of patients in whom this procedure was performed, though urgent coronary angiography was indicated but not performed in nine patients admitted to nontertiary hospitals.
Median delay in performing urgent catheterization was shorter in the tertiary hospital than in the rest (within 24 h vs. four days, respectively, p < 0.0001). The number of catheterizations performed within two (p = 0.0002) and 28 days (p < 0.0001) after onset of UA symptoms was higher in the tertiary hospital (Table 3).
Among the 88 patients with urgent indication for coronary angiogram, the proportions of patients finally receiving PTCA or CABG were similar in both groups, as were median delays in performing urgent revascularization procedures. The proportion of patients meeting urgent criteria for PTCA differed between tertiary and nontertiary hospitals as occurred in urgent angiography although with less significance. The proportions of elective PTCA and CABG were higher in the tertiary hospitals (Table 3).
Both 28-day and six-month mortality and readmission rates per group are shown in Table 4. No patient was lost to follow-up. Overall intergroup differences in mortality or readmission rate were not statistically significant. Survival curves in both groups showed no statistically significant difference (Fig. 1). Furthermore, no differences existed in the causes of readmission, including AMI and new episodes of UA (Table 4).
Logistic regression models adjusted for differences between groups A and B (Table 5)showed that hospital type was not an independent risk factor for six-month mortality or readmission (RR for group B 1.23, 95% CI: 0.57 to 2.67). No differences between groups were found when only six-month mortality was considered as a dependent variable (RR for group B was 1.27, 95% CI: 0.86 to 1.88). Statistically significant interaction terms were identified in none of these models (Table 5).
Results of the present study show that the use of tertiary procedures in patients with UA without previous AMI was higher in the tertiary hospital, but such a difference did not translate to better outcome, not only in terms of mortality or all-cause readmission rates but also when AMI or new episodes of UA were considered alone. The difference in resources use rates in AMI patients has been reported in several studies comparing tertiary versus nontertiary hospitals (4,5)and Canadian versus U.S. hospitals (8)with no differences found in outcome. Studies focusing on this issue in UA patients are sparse. One study (9)comparing Canadian with U.S. hospitals in UA and non–Q-wave AMI patients showed a similar use rate in both countries, which contrasts with findings in AMI patient studies. Another study, made in The Netherlands (10)in UA patients, showed, as in ours, a higher use of coronary angiography in the tertiary hospital, also without differences in outcome.
Interestingly, no differences were found in the proportion of patients undergoing urgent diagnostic or therapeutic procedures, though longer delays were encountered in patients admitted to nontertiary hospitals. Furthermore, in our study the proportion of patients considered to have an indication of urgent coronary angiography was lower in the tertiary hospital, probably in part as a consequence of more elective procedures being performed within the first 48 h of admission to the center.
These results raise the question of whether differences in use reflect an excess in the number of procedures performed in the tertiary hospital or if under-use occurs in nontertiary centers. In any event, the absence of differences in outcome in our study suggests that the more selective procedure use rate in nontertiary hospitals provides appropriate management of UA patients. Furthermore, invasive management has not proved better than clinically guided management in the literature (11–13), which may account for a substantial amount of patients with acute coronary syndromes undergoing coronary angiography and revascularization without clear indication (14). Indeed, large geographic variations in procedure use rate have been described, which indicates that their use is influenced by some nonobjective medical reasons (8,14,15). In addition to availability, financial incentives and patient demand have been suggested as factors influencing variations in the use of tertiary procedures in coronary patients (8,14). Hospitals in the Spanish National Health System do not provide financial incentives for physicians to perform procedures. It is therefore highly unlikely that financial reasons accounted for the differences observed in our study.
A similar assessment to that presented here for UA has been previously published for AMI patients in the same setting (5). Therefore, comparison of tertiary resources use rates of both coronary events may be undertaken straightforwardly in the same scenario. To our knowledge, no previous work has been able to compare invasive-procedure use rates of both Q-wave AMI and UA patients in tertiary and nontertiary hospitals. Procedures were used more frequently in UA than in first AMI patients in both types of hospitals (Fig. 2). Different clinical characteristics of both coronary syndromes and the idea of preventing AMI may have accounted for this phenomenon. In contrast, procedures were more frequently used in the tertiary hospital in both coronary acute syndromes. As shown in Figure 2, the magnitude of the use-rate difference between the two types of hospitals was less marked in UA than in first AMI patients. In our study, nontertiary centers were clearly prone to using more resources in UA. This is why the differences with the tertiary hospital were less pronounced in UA patients than in first AMI patients, in whom management in nontertiary centers is much more conservative.
Study characteristics and limitations
The observation period was extended to six months, when most outcomes related to the index event probably would already have occurred; it is difficult to project whether the survival curves would diverge in an extension of the follow-up.
The present study did not address other issues related to on-site unavailability of tertiary care procedures such as inconvenience and distress for patients transferred between hospitals and for patients’ relatives. Mortality and coronary events may be too rough an index to evaluate the outcome of UA patients; other end points such as quality of life might provide complementary information in patients with several heart diseases (16). Some studies show an improvement in quality of life with a higher use rate of invasive procedures in patients with AMI (17). This issue was not addressed in our study.
The following measures were taken to ensure efficient case-mix control: only UA patients without previous myocardial infarction were included, exclusions were recorded and their causes justified, and statistical adjustment for case-mix (i.e., differences in severity and co-morbidity between the two hospital types) was used. To prevent physician-dependent outcomes, PTCA or CABG after discharge were not used as end points.
In conclusion, the results of the present study suggest that the use rate of coronary angiography and revascularization procedures in UA patients with no previous AMI is higher in tertiary than in nontertiary hospitals, but the more selective use in nontertiary facilities does not imply worse outcome. The use rate of these procedures is higher in UA than in first AMI patients in both types of hospitals, particularly in the nontertiary centers.
The authors appreciate the English version of the manuscript prepared by Ms Christine O’Hara.
☆ This project was funded by grant 92/0009 from the Fondo de Investigación Sanitaria and by the Generalitat de Catalunya grant CIRIT/SGR 9700218.
- acute myocardial infarction
- confidence interval
- coronary artery bypass grafting
- electrocardiogram, electrocardiographic
- percutaneous transluminal coronary angioplasty
- Recursos Empleados en el Sı́ndrome Coronario Agudo y Tiempos de Espera
- unstable angina
- Received December 3, 1998.
- Revision received June 24, 1999.
- Accepted August 27, 1999.
- American College of Cardiology
- Théroux P,
- Fuster V
- RESCATE Investigators,
- Marrugat J,
- Sanz G,
- Masià R,
- et al.
- Hosmer D.W,
- Lemeshow S
- Van Miltenburg-van Zijl A.J.M,
- Simoons M.L,
- Bossuyt P.M.M,
- et al.
- TIMI-IIIB Investigators
- Vernon Anderson H,
- Cannon C,
- Stone P,
- et al.
- Veterans Affairs Non-Q-Wave Infarction Strategies in Hospital (VANQWISH Trial Investigators),
- Boden W.E,
- O’Rourke R.A,
- Crawford M.H,
- et al.