- Email: [email protected]
Early discharge after acute myocardial infarction in the current clinical practice. Community data from the AMI-Florence Registry, Italy
International Journal of Cardiology 114 (2007) 57 – 63 www.elsevier.com/locate/ijcard
Early discharge after acute myocardial infarction in the current clinical practice. Community data from the AMI-Florence Registry, Italy Alessandro Barchielli a,*, Daniela Balzi a, Niccolo` Marchionni b, Nazario Carrabba c, Massimo Margheri d, Giovanni M. Santoro e, Iacopo Olivotto f, Eva Buiatti g the AMI-Florence Working Group1 a Epidemiology Unit, Local Health Unit Florence, Via di San Salvi 12, 50135 Florence, Italy Department of Critical Care Medicine and Surgery, Unit of Gerontology and Geriatric Medicine, University of Florence, and Careggi Hospital, Florence, Italy c Cardiology Unit 1, Careggi Hospital, Florence, Italy Department of Critical Care Medicine and Surgery, Unit of Internal Medicine and Cardiology, University of Florence, and Careggi Hospital, Florence, Italy e Cardiology Unit, Nuovo San Giovanni di Dio Hospital, Local Health Unit Florence, Florence, Italy f Emergency Department, Careggi Hospital, Florence, Italy g Epidemiology Unit, Regional Agency for Public Health of Tuscany, Florence, Italy b
d
Received 18 July 2005; received in revised form 27 December 2005; accepted 3 January 2006 Available online 18 May 2006
Abstract Background: Clinical trials suggested feasibility and safety of early discharge after ST-segment elevation acute myocardial infarction (STEMI) for selected patients. Current United States and European guidelines recommend early discharge for uncomplicated AMI. The present study was aimed to assess early discharge in the current clinical practice for STEMI patients. Methods: Patients enrolled in the AMI-Florence Registry (Italy), a prospective, observational, population-based study performed in the early 2000s, were analysed. The proportion of STEMI patients eligible for early discharge and of those actually discharged early, patient features influencing early discharge and outcomes (mortality, reinfarction or urgent revascularization during the first year of follow-up) were assessed. Results: Of 751 STEMI patients, 59% were classified as eligible for early discharge, according to previously established criteria. Among patients considered eligible, those actually discharged early dropped from 33.1% in age < 70 years to 15.9% in age 70 – 79 years and to 11.7% in age 80+ years. Of eligible patients, 26% were actually discharged within 4 days. Age 70+ years (reference: 69 years or younger; 70 – 79 years: OR: 0.40, 95% CI: 0.22 – 0.73; 80+ years: OR: 0.33, 95% CI: 0.14 – 0.76) and diabetes (OR: 0.48, 95% CI: 0.24 – 0.98) were negative independent predictors; and coronary reperfusion (OR: 2.92, 95% CI: 1.54 – 5.53) or firstly admitted to teaching hospital (OR: 1.68, 95% CI: 1.03 – 2.74) were positive predictors, of the multivariate probability of being early discharged. Among patients eligible for early discharge, 1-year and 3-year mortality did not differ significantly between patients actually, and not, early discharged. After 1 year of followup, no deaths were observed among patients early discharged.
* Corresponding author. Tel.: +39 556263373; fax: +39 556263375. E-mail address: [email protected] (A. Barchielli). 1 AMI-Florence Working Group: Careggi Hospital: G.M. Santoro, N. Carrabba (Cardiology Unit 1), G. Santoro, G. Corti (Cardiology Unit 2), M. Margheri, S. Valente (Department of Critical Care Medicine and Surgery, Unit of Internal Medicine and Cardiology), F. Ferrante (Medicine Unit 1), V. Verdiani (Medicine Unit 2), I. Olivotto (Medicine Unit 3), N. Marchionni, M. Monami (Department of Critical Care Medicine and Surgery, Unit of Gerontology and Geriatric Medicine); Local Health Unit Florence: A. Barchielli, D. Balzi, P. Naldoni, C. Melani (Epidemiology Unit), L. Fratoni (Emergency Ambulance Transport 118 Service); Santa Maria Annunziata Hospital: A. Fantini (Coronary Care Unit), M. Torri (Medicine Unit 1), G. Regoli, C. Mugnaini (Medicine Unit 2); Santa Maria Nuova Hospital: M.C. Landini (Coronary Care Unit), M. Granelli (Medicine Unit 1); Nuovo San Giovanni di Dio Hospital: M. Filice (Coronary Care Unit), P. Fabiani (Medicine Unit 1); Mugello Hospital: F. Miglietta (Intensive Therapy Unit), L. Scarti, I. Berni (Medicine Unit); Serristori Hospital: G. Fabrizi de Biani, F.A. Tarmun (Medicine Unit); Regional Agency for Public Health of Tuscany: E. Buiatti, S. Arniani, S. Bartolacci (Epidemiology Unit). 0167-5273/$ - see front matter D 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijcard.2006.01.006
58
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
Conclusion: This study confirms, in a setting of current clinical practice, the reliability of criteria for identifying patients eligible for early discharge. Besides, the results confirmed the safety of this practice in selected patients. About 1/4 of eligible patients are actually early discharged, confirming the existence of important opportunities to improve the efficiency in STEMI management. D 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Acute myocardial infarction; Population-based registry; Prognosis; Early discharge
1. Introduction Improvement in medical care (e.g. recognition of the beneficial effects of early mobilization; increased use of coronary reperfusive therapy), and increasing pressure toward cost containment in health care, have resulted in a progressive shortening of hospital stay after acute myocardial infarction (AMI) [1]. In early 1990s, some clinical trials suggested the feasibility and the safety of early discharge after AMI (3– 4 days), at least for selected subgroups of patients [2– 4]. Simple clinical characteristics have been defined to identify cases of ‘‘uncomplicated infarction’’: these include absence of reinfarction, myocardial ischaemia, stroke, hearth failure, need for by-pass surgery, intra-aortic balloon pumping, emergency catheterization, cardioversion or defibrillation in the first four in-hospital days [5]. When applied to patients in the GUSTO-I trial, these criteria could identify patients at very low risk, who were considered eligible for early discharge from hospital [5]. In keeping with these criteria, current US and European guidelines recommend an early discharge for patients with uncomplicated AMI [6,7]. Recently, prognostic scores for ST-segment elevation myocardial infarction (STEMI) treated with primary angioplasty have been proposed [8,9]. In particular, the score proposed by De Luca et al. reliably identifies a large group of patients at very low risk, who may safely discharged early after primary angioplasty [8]. Cost/effectiveness analyses suggested that, in relation to conventional standards, prolonged hospital stay after thrombolysis or primary angioplasty of patients with uncomplicated myocardial infarction is ‘‘economically unattractive’’ [8,10] and an early discharge strategy for low-risk patients results in a substantial cost saving [2,11,12]. Despite this, Kaul et al. [13] recently have pointed out that, in the setting of some international clinical trials, no more than 40% of patients eligible for early discharge were actually discharged early. In particular, the rate of early discharge of eligible patients was especially low in the European Countries included in the analysis (Belgium, France, Germany, Spain and Poland). The Florence Acute Myocardial Infarction (AMI-Florence) Registry is a population-based, observational study carried out in the Florence health district (Tuscany region, central Italy) [14], enrolling STEMI. In this area, primary percutaneous coronary intervention (PCI) is the most widely preferred myocardial reperfusive treatment. The present
analysis of the AMI-Florence dataset was aimed at identifying in the ‘‘real world’’ patients eligible for early discharge on the basis of criteria previously specified [5] and to determine the proportion of those actually discharged early. Furthermore, this study analysed the clinical factors that influence early discharge in the current clinical practice and it verified outcomes in relation to the length of hospital stay.
2. Materials and methods 2.1. Registry design and setting The structure and setting of the AMI-Florence Registry have been detailed elsewhere [14]. Briefly, the Florence health district (about 800,000 inhabitants) comprises five community hospitals and one teaching hospital, the latter implementing high-volume program for primary PCI, operative 24 h/day, 7 days/week. Patients considered eligible for coronary reperfusion are treated with primary PCI or fibrinolysis. Patients admitted to the teaching hospital are treated with primary PCI, while those admitted to community hospitals, are usually transferred to the teaching hospital for primary PCI or, less frequently, receive fibrinolytic therapy onsite [14]. The AMI-Florence Registry was a prospective, observational, population-based study, which included all cases of AMI arriving alive to one out of six hospitals in the Florence health district, with no exclusion criteria (recruitment period: 2000 – 2001). Patients with a diagnosis of suspected AMI were prospectively screened for eligibility and enrolled in the study if AMI was confirmed by at least two of the following criteria: typical chest pain lasting > 30 min, rise in serum creatine phosphokinase activity at least twice above the normal upper limit within 72 h of symptom onset and evidence of AMI on the first ECG (ST-segment elevation 0.1 mV in two or more adjacent leads or new onset left bundlebranch block). Information on demographics, medical history, clinical and ECG features of AMI, time delay between onset of symptoms and hospital admission, treatment and outcome during hospitalization, was collected using standard case-report forms. In our country, at the time of patients enrolment, ethics committee approval and informed consent from patients were not required for observational studies.
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
59
surgery or died within 4 days from hospital admission, were excluded from the present analysis (169/920).
2.2. Patients categorization Using criteria similar to those previously established in the setting of clinical trials [5,13], STEMI cases enrolled in the AMI-Florence Registry (920 patients) were defined eligible for early discharge (within 4 days of admission) when none of the following events occurred during the hospital stay: reinfarction, myocardial ischaemia, stroke, shock, heart failure (Killip class > 1), by-pass surgery, intraaortic balloon pumping, emergency cardiac catheterization (for assessment of sign or symptoms of clinical instability), or need for cardioversion or defibrillation. Patients who had a history of stroke, those who recently had undergone major
2.3. Follow-up Major adverse events, including nonfatal reinfarction, urgent revascularization and death for all causes, were evaluated during the first year after the index AMI. To this purpose, subsequent hospitalizations were identified through the Regional Hospital Discharges database, recording all admissions to regional public and private hospitals. The original clinical records of re-hospitalized patients were checked to identify the occurrence of non-fatal reinfarction
Table 1 Patients eligible for early discharge: demographic and clinical characteristics by length of hospital stay Variable
Demographic Female gender Age (mean) Medical history Previous myocardial infarction Angina, onset >1 month Angina, onset 1 month CABG PCI Atrial fibrillation, paroxysmal Atrial fibrillation, chronic Conduction disorders Congestive heart failure Depression Peripheral artery diseases Renal failure, serum creatinine 1.5 – 2.5 mg/dl Renal failure, serum creatinine >2.5 mg/dl Chronic obstructive pulmonary disease Gastrointestinal diseases Cancer, onset 5 years Cancer, onset >5 years Anaemia Diabetes Hypertension Dyslipidaemia Smokers, current Smokers, former Hospital admission Directly admitted to hospital with PCI facilities Arrival modalities: referred by mobile care unit Admitted during office hours Time symptom/hospital arrival 6h Time hospital arrival/reperfusive therapy (median) Clinical presentation, AMI characteristics, treatment CPK peak 72 h (median) AMI anterior location, Q-waves AMI other location, Q-waves AMI any location, non-Q-waves Reperfusive therapy (primary PCI or fibrinolysis) Fibrinolysis Primary PCI Stenting rate (PCI patients) TIMI 3 grade flow after PCI
Length of hospital stay
p-value
4 days (115 cases)
5+ days (327 cases)
23.5% 60.5 years
29.1% 66.9 years
7.8% 13.9% 24.4% 0.0% 5.2% 0.9% 2.6% 1.7% 0.0% 3.5% 6.1% 0.9% 0.0% 6.1% 20.9% 0.9% 3.5% 0.9% 9.6% 41.7% 31.3% 51.3% 17.4%
14.7% 15.9% 18.4% 2.1% 4.3% 4.0% 2.8% 6.1% 1.2% 3.4% 8.3% 4.3% 2.1% 6.7% 19.9% 1.8% 4.0% 4.0% 20.0% 49.2% 30.0% 35.2% 20.2%
<0.008
66.1% 91.3% 67.8% 82.2% 30 min
43.4% 85.9% 61.5% 75.8% 42 min
<0.001 0.12 0.22 0.03 0.40
1280 U/l 25.2% 60.9% 13.9% 87.0% 0.0 87.0% 96.0% 96.0%
1255 U/l 28.8% 52.3% 19.0% 61.8% 6.4% 55.4% 95.0% 95.0%
PCI: percutaneous coronary intervention, CABG: coronary artery bypass graft, AMI: acute myocardial infarction, CK: creatine phosphokinase.
0.25 <0.001 0.059 0.37 0.11 0.68 0.26 0.063 0.23 0.95 0.45 0.059 0.81 0.82 0.75 0.10 0.012 0.17 0.79
0.82
<0.25 <0.001 <0.001 0.71 0.31
60
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
or need for urgent coronary revascularization. Reinfarction was diagnosed in the presence of the above reported criteria for the diagnosis of the index AMI. Coronary revascularization, performed with percutaneous intervention or surgery, was considered urgent when not performed on elective basis. Mortality was checked by consulting the Registry Office of the municipalities of residence. 2.4. Statistical analysis Differences across patients sub-groups were evaluated with the v 2 test for categorical variables and with the Student’s t-test for continuous variables. All tests were twosided and statistical significance was defined as a p < 0.05. Stepwise multivariate regression analyses (forward method, with p < 0.10 for entrance into and p > 0.15 for removal from the model) were performed to identify factors associated with early discharge and prognosis. Multivariate logistic regression was used to identify among the demographic and clinical characteristics of patients (reported in Table 1) the independent determinants of early discharge. The goodnessof-fit of the model was checked through the HosmerLemeshow test. Odds ratios (OR) and 95% confidence intervals (95% CI) were calculated. In addition, Cox regression was used to assess the relationship between early discharge and prognosis after adjustment for clinical and demographic variables with an independent prognostic effect. To this purpose, hazard ratios (HR) and 95% confidence intervals (95% CI) were calculated. The proportional hazard assumption was checked through the test proposed by Grambsch and Therneau, and the goodness-offit of the model through the test proposed by May and Hosmer [15]. Analyses were carried out using STATA statistical package (version 6.0, Stata Corporation, College Station, Texas, 1999).
3. Results 3.1. Patients characteristics Of 751 patients included in the present analysis (median age: 70 years, interquartile range: 60 –79), 442 (59%) were classified eligible for early discharge and 134 (17.8%) discharged alive within 4 days (Fig. 1). Because of the community composition of Florence health district, many patients enrolled in the Registry were elderly. In particular, cases considered not eligible were significantly older than those considered eligible for early discharge (mean age: 76.4 vs. 65.2 years, p < 0.001), the percentage of eligible decreasing from 75.8% in age < 70 years to 48.2% in age 70 –79 years and to 36.1% in age 80+ years. Besides, not eligible patients were more frequently females (35.6% vs. 27.6%, p = 0.020), affected by previous cardiovascular (i.e.: congestive heart failure: 7.0% vs. 0.3%, p < 0.001; angina of recent onset: 24.6% vs. 15.4%, p = 0.004; atrial fibrillation:
Fig. 1. Patients enrolled in AMI-Florence Registry and included in the analysis, by subgroups.
11.0% vs. 5.9%, p = 0.001; conduction disorder: 11.7% vs. 5.0%, p = 0.001; peripheral artery disease: 14.2% vs. 7.7%, p = 0.004) and non-cardiovascular diseases (i.e.: depression: 7.4% vs. 3.4%, p = 0.013; renal failure: 10.0% vs. 5.0%, p = 0.026; chronic obstructive pulmonary disease: 13.3% vs. 6.6%, p = 0.002; diabetes: 27% vs. 16%, p = 0.008), less frequently admitted directly to the teaching hospital with PCI facilities (33.3% vs. 49.3%, p < 0.001), with a higher proportion of anterior STEMI (44.0% vs. 27.8%, p < 0.001), less frequently treated with reperfusion (48.9% vs. 68.3%, p < 0.001) and, as expected, with a worse prognosis (i.e.: 1-year mortality: 24.0% vs. 4.3%, p < 0.001). Among patients considered eligible for early discharge, 115 (26.0%) were actually early discharged. Patients discharged early were significantly younger and less frequently affected by diabetes (Table 1). Besides, patients discharged early showed a lower prevalence (borderline significant) of previous myocardial infarction, conduction disorders and renal failure. On the other hand, more frequently, they were admitted directly to the hospital with PCI facilities and treated with reperfusive therapy. Interestingly, coronary reperfusion (with primary PCI being the most common strategy: 93.0%, 281/302) was significantly more frequent among cases directly admitted to the teaching hospital with PCI facilities than to community hospitals (88.1% vs. 49.1%, p < 0.001). Besides, patients directly admitted to the hospital with PCI facilities were significantly less affected by conduction disorder (2.8% vs. 7.1%, p = 0.034) and more frequently admitted within 6 h from symptom onset (85.8% vs. 71.0%, p = 0.001), and borderline significantly younger (mean age: 64 vs. 66 years, p = 0.071). Among patients considered eligible, those actually discharged early dropped from 33.1% (91/275) in age < 70 years to 15.9% (17/107) in age 70 –79 years and to 11.7% (7/60) in age 80+ years. A large part of patients eligible for early discharge arrived at hospital within 6 h of symptoms of onset (0 – 3 h: 56.6%, 4 – 6 h: 21.7%), the remaining 21.7% arriving 7 h or
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
61
later. Patients eligible and actually discharged early decreased from 28.8% when admitted within 3 h to 27.1% when admitted within 4 –6 h and to 17.7% when admitted 7 h or later from symptoms onset. 3.2. Length of hospital stay of patients eligible for early discharge By definition, the length of stay during the first hospital admission was significantly shorter in patients discharged early compared to those late discharged (respectively: 3.4 and 9.1 days), with a mean difference of about 6 days, largely explaining the difference observed in the total length of hospital stay during the first year after the index AMI (respectively: 9.4 and 16.6 days) (Table 2). Patients discharged early during the first hospitalization showed a shorter length of stay for the following hospital admissions during the 1-year follow-up, but this difference was less relevant (about 1.5 days; respectively: 6.0 and 7.4 days). The length of hospital stay (Fig. 2) differed significantly between patients firstly admitted to teaching and community hospitals (mean: 6.8 vs. 8.5 days, p = 0.002) and, more markedly, between patients treated and not treated with reperfusive treatment (mean: 7.0 vs. 9.0 days, p < 0.001). The difference across the hospitals of first admission was not significant when the comparison was stratified by treatment.
Fig. 2. Patients eligible for early discharge: length of hospital stay stratified by hospital of first admission and reperfusive treatment (TH: teaching hospital, CH: community hospitals, R+: reperfusive treatment, R : conservative treatment).
Among patients aged <70 years, the multivariate logistic regression analysis showed that the probability to be discharged early was borderline significantly lower for patients with diabetes (OR: 0.49, 95% CI: 0.22 – 1.10, p = 0.083) and significantly higher for those treated with reperfusion therapy (OR: 3.74, 95% CI: 1.84 – 7.60, p < 0.001). Among those aged 70+ years, the probability to be discharged early was significantly higher among those directly admitted to the teaching hospital (OR: 2.78, 95% CI: 1.09– 7.11, p = 0.033).
3.3. Factors influencing early discharge at the multivariate analysis
3.4. Outcomes
At the multivariate logistic regression analysis, the probability to be discharged early was significantly lower among patients aged 70 years and older (reference: 69 years or younger; 70 – 79 years: OR: 0.40, 95% CI: 0.22 –0.73, p = 0.003; 80 years or older: OR: 0.33, 95% CI: 0.14 –0.76, p = 0.003) or with diabetes (OR: 0.48, 95% CI: 0.24 –0.98, p = 0.043), and higher among those treated with reperfusion therapy (OR: 2.92, 95% CI: 1.54 – 5.53, p = 0.001) or directly admitted to the teaching hospital (OR: 1.68, 95% CI: 1.03– 2.74, p = 0.038).
After 1 year of follow-up, patients discharged early had not significantly different rates of reinfarction and of the cumulative incidence of major adverse events than those discharged late (Table 2). On the other hand, whereas urgent revascularization was performed more often among patients discharged early (difference borderline significant), mortality was significantly lower for patients discharged early. In the multivariate Cox analysis, the difference in urgent revascularization between the two groups of patients was not significant ( p = 0.14). No deaths in the discharged early
Table 2 Patients eligible for early discharge: days spent in hospital and outcomes by length of hospital stay Variable
All patients (442)
Length of hospital stay 4 days (115 cases)
Length of hospital stay (mean) First admission Other admissions in the first year of follow-up Total admissions in the first year of follow-up Outcomes (%) 1-year reinfarction 1-year urgent revascularisation 1-year mortality 1-year major adverse eventsa 3-year mortality a
Death, reinfarction or urgent revascularisation.
7.6 7.1 14.7 4.1 13.8 4.1 20.8 9.5
3.4 days 6.0 days 9.4 days 3.5% 19.1% 0.0% 22.6% 1.7%
p-value
5+ days (327 cases) 9.1 days 7.4 days 16.6 days 4.3% 11.9% 5.5% 20.2% 12.2%
<0.001 0.31 <0.001 0.71 0.054 0.010 0.48 0.001
62
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
group were observed in the first year of follow-up; therefore, the Cox analyses were only performed after 3 years of follow-up. Whereas in the univariate Cox analysis, patients discharged early showed a significant lower risk of death (HR: 0.13, 95% CI: 0.03 –0.56, p = 0.006), at the multivariate Cox analysis, after adjusting for the other variables with an independent prognostic effect (age, reperfusive therapy and medical history of anaemia, renal failure or depression), the difference in mortality was no more significant (HR: 0.42, 95% CI: 0.10– 1.82, p = 0.25). It is noteworthy that, among patients aged <70 years and discharged with 4 days, no deaths were observed after 3 years of follow-up. Only 1.7% (2/115) of the patients discharged early had an adverse event (reinfarction or urgent revascularization) after discharge and within the ninth day from the hospital admission (the mean number of days of hospital stay for patients discharged later).
4. Discussion The results of this population-based study carried out in the early 2000s showed that a large proportion (59%) of STEMI patients fulfilled criteria previously established by Newby et al. [5] for being candidates to early hospital discharge. In our study, the proportion of patients eligible for early discharge is similar to those reported in some clinical trials for other European countries, which ranged from 57% in France to 70% in Poland in the ASSENT-2 trial, carried out in the late 1990s [13]. However, in these countries, only 2% or less of eligible patients were actually discharged early. This proportion was very low, particularly when compared to the about 40% of early discharges that has been reported in USA and New Zealand [13]. In our study, about 26% of eligible patients were actually early discharged, a proportion greater than those previously reported in other European countries. Some issues must be discussed which are likely to account, at least in part, for this large difference. In Italy, where the National Health System covers hospital care without any charge for patients, a diagnosis-related-groups (DRGs) payment system for hospital reimbursement was introduced since the middle 1990s. As a consequence, substantial efforts have been made to improve performance and efficiency in hospital management, resulting in a general shortening in the length of hospital stay. Furthermore, our data was collected some years after the ASSENT-2 trial, when spreading of updated guidelines in the clinical practice might have contributed to reduce the length of hospital stay after uncomplicated AMI. Nevertheless, about 2/3 of eligible patients were discharged after 5 or more days in hospital suggesting that, as previously observed [13], important opportunities still remain to improve efficiency in AMI care. Our data were derived from an observational study that enrolled all patients with STEMI from the target population
who were hospitalized during the recruitment period, without selection bias. This is an important reason of interest for our results. Indeed, whereas previous analyses were referred to the selected setting of clinical trials [13], our study was aimed at analysing data collected in the perspective of current clinical practice and, hence, referring to the ‘‘real world’’. In particular, older patients are often underrepresented in clinical trials [16], whereas in the ‘‘real world’’ a large proportion of AMI patients are older [17]. For example, the median age is 61 years in the ASSENT-2 trial [13], compared to 70 years in study (where about 52% of patients were aged 70 years). In our series, patients eligible for early discharge, according to Newby et al. [5] criteria, dropped from 76.8% in age < 70 years to 36.1% in age 80+ years, and eligible patients actually discharged early from 33% (91/275) to 11.7% (7/60). The multivariate analysis showed that ‘‘eligible’’ patients had a significantly lower probability of being actually discharged early when aged 70 years or affected by diabetes. As expected in the clinical practice, some characteristics of AMI patients that are not effectively captured using standardized criteria may influence physicians’ final decision about hospital discharge. This finding suggests that the presence of noncardiac factors (such as diabetes or of other important comorbidities) should be considered to improve the reliability of criteria for identification of low-risk patients eligible for early discharge. Nevertheless, even after exclusion from analysis of cases with diabetes, the proportion of eligible patients who were actually discharged early moderately increased to 28%, reinforcing the view that the efficiency of management of AMI can still be improved. On the other hand, the probability of early discharge was higher in patients directly admitted to the hospital with PCI facilities or treated with reperfusion therapy. From a general point of view, both factors can influence early discharge practice. In fact, some hospital may have a more pressing interest in shortening the length of hospital stay, so that it can increase the number of hospitalizations. On the other hand, coronary reperfusion (and, in particular, primary PCI) is currently recognized as the most appropriate and beneficial treatment for AMI [6,7,18] and use of beneficial treatments is likely to reduce the length of hospital stay [19,20]. In particular, regardless of clinical risk status, the success of PCI has been proved to be a strong independent predictor of 30-day major adverse cardiac events, and an early discharge policy for patients with successful PCI and less than three high-risk features has been suggested [19]. In our dataset, admission hospital and reperfusive treatment have a certain degree of colinearity. In fact, about 88% of patients directly admitted to the teaching hospital were treated with reperfusive treatment compared to 49% in community hospitals. In a previous analysis of the whole series, the hospital of admission was the strongest independent predictor for subsequent PCI, the higher perception of indication to reperfusion by attending physician in hospital with PCI facilities likely explain this finding [14]. In the present
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
analysis, the result of multivariate analysis suggested that both reperfusive therapy and type of hospital of first admission were independent predictors of early discharge. Interestingly, the proportion of patients eligible for early discharge and of those actually discharged early was higher among patients early presented, suggesting an influence of time between symptoms onset and hospital arrival (and therefore to proper AMI treatments) also on the length of hospital stay. Hypothesizing the feasibility to apply an early discharge policy to all eligible patients, about 748 days spent in hospital could be potentially saved in age <70 years, 428 in age 70– 79 years and 261 in age 80+ years (respectively 22%, 18% and 8% of the total number of days spent in hospital during the first hospitalization). Mortality rates observed in this study confirmed, also at the level of current clinical practice, the reliability and the safety of criteria that originally have been determined in the setting of clinical trials for identifying low-risk patients eligible for early discharge [5,13]. Indeed, 1-year mortality of eligible patients was very low (about 4% for all subjects), and 1-year mortality and other clinically relevant adverse events (such as reinfarction or need for urgent PCI) did not significantly differ, within the group of eligible patients, between those who were and those who were not actually discharged early. It is noteworthy that, in this group of patients, no deaths were observed among discharged early patients after 1 year of follow up and that, after 3 years of follow-up, only 2 (1.7%) patients aged > 70 years died, whereas no deaths were reported among patients aged 70 years or less. In conclusion, our data confirm, in the current clinical practice, the reliability of criteria for identification of lowrisk patients eligible for early discharge. Besides, the results confirm the safety of this practice in selected patients. Despite this evidence, only about 1/4 of patients eligible to early discharge are actually discharged within 4 days of admission, confirming the existence of important opportunities to improve efficiency in AMI care.
References [1] Spencer FA, Lessard D, Gore J, et al. Declining length of hospital stay for acute myocardial infarction and postdischarge outcomes. Arch Intern Med 2004;164:733 – 40. [2] Topol EJ, Burek K, O’Neill WW, et al. A randomised trial of hospital discharge three days after myocardial infarction in the era of reperfusion. N Engl J Med 1988;318:1083 – 8. [3] Mark DB, Sigmon K, Topol EJ, et al. Identification of acute myocardial infarction patients suitable for early hospital discharge after aggressive interventional therapy. Results from the Thrombolysis and Angioplasty in Acute Myocardial Infarction Registry. Circulation 1991;83:1186 – 93.
63
[4] Sanz G, Betriu A, Oller G, et al. Feasibility of early discharge after acute Q wave myocardial infarction in patients not receiving thrombolytic treatment. J Am Coll Cardiol 1993;22:1795 – 801. [5] Newby LK, Califf RM, Guerci A, et al. Early discharge in the thrombolytic era: an analysis of criteria for uncomplicated infarction from the Global Utilization of Strepokinase and t-PA for Occluded Coronary Arteries (GUSTO) trial. J Am Coll Cardiol 1996;27:625 – 32. [6] Ryan TJ, Antman EM, Brooks NH, et al. 1999 update: ACC/AHA guidelines for the management of patients with acute myocardial infarction: executive summary and recommendations. Circulation 1999;100:1016 – 30. [7] Van de Werf F, Ardissino D, Betriu A, et al. Management of acute myocardial infarction in patients presenting with St-segment elevation. Eur Heart J 2003;28:28 – 66. [8] De Luca G, Suryapranata H, van’t Hof AW, et al. Prognostic assessment of patients with acute myocardial infarction treated with primary angioplasty. Circulation 2004;109:2737 – 43. [9] Morrow DA, Antman EM, Charlesworth A, et al. TIMI risk score for ST-elevation myocardial infarction: a convenient, bedside, clinical core for risk assessment at presentation. Circulation 2000; 102:2031 – 7. [10] Newby LK, Eisenstein EL, Califf RM, et al. Cost effectiveness of early discharge after uncomplicated acute myocardial infarction. N Engl J Med 2000;342:749 – 55. [11] Grines CL, Marsalese DL, Brodie B, et al. Safety and costeffectiveness of early discharge after primary angioplasty in low risk patients with acute myocardial infarction. J Am Coll Cardiol 1998;31: 967 – 72. [12] Bartholomew BA, Harjai KJ, Grines CL, et al. Variation in hospital length of stay in patients with acute myocardial infarction undergoing primary angioplasty and the need to change the diagnostic-related groups system. Am J Cardiol 2003;93:830 – 3. [13] Kaul P, Newby LK, Fu Y, et al. International differences in evolution of early discharge after acute myocardial infarction. Lancet 2004;363: 511 – 7. [14] Buiatti E, Barchielli A, Marchionni N, et al. Determinants of treatment strategies and survival in acute myocardial infarction: a population based study in the Florence district, Italy: results of the Acute Myocardial Infarction Florence Registry (AMI-Florence). Eur Heart J 2003;24:1195 – 203. [15] May S, Hosmer D. A simplified method of calculating an overall goodness-of-fit test for the Cox proportional hazard model. Lifetime Data Anal 1998;4:109 – 20. [16] Lee PY, Alexander KP, Hammill BG, et al. Representation of elderly persons and women in published randomized trials of acute coronary syndromes. JAMA 2001;286:708 – 13. [17] Parmley WW. Do we practice geriatric cardiology? J Am Coll Cardiol 1997;29:217 – 8. [18] Aversano T, Aversano LT, Passamani E, et al. Thrombolytic therapy vs. primary percutaneous coronary intervention for myocardial infarction in patients presenting to hospitals without on-site cardiac surgery. JAMA 2002;287:1943 – 51. [19] Heggunje PS, Harjai KJ, Stone GW, et al. Procedural success versus clinical risk status in determining discharge of patients after primary angioplasty for acute myocardial infarction. J Am Coll Cardiol 2004; 44:1400 – 7. [20] Dirksen MT, Ronner E, Laarman GJ, et al. Early discharge is feasible following primary percutaneous coronary intervention with transradial stent implantation under platelet glycoprotein IIb/IIIa receptor blockade. Results of the AGGRASTENT trial. J Invasive Cardiol 2005;17: 512 – 7.
Early discharge after acute myocardial infarction in the current clinical practice. Community data from the AMI-Florence Registry, Italy Alessandro Barchielli a,*, Daniela Balzi a, Niccolo` Marchionni b, Nazario Carrabba c, Massimo Margheri d, Giovanni M. Santoro e, Iacopo Olivotto f, Eva Buiatti g the AMI-Florence Working Group1 a Epidemiology Unit, Local Health Unit Florence, Via di San Salvi 12, 50135 Florence, Italy Department of Critical Care Medicine and Surgery, Unit of Gerontology and Geriatric Medicine, University of Florence, and Careggi Hospital, Florence, Italy c Cardiology Unit 1, Careggi Hospital, Florence, Italy Department of Critical Care Medicine and Surgery, Unit of Internal Medicine and Cardiology, University of Florence, and Careggi Hospital, Florence, Italy e Cardiology Unit, Nuovo San Giovanni di Dio Hospital, Local Health Unit Florence, Florence, Italy f Emergency Department, Careggi Hospital, Florence, Italy g Epidemiology Unit, Regional Agency for Public Health of Tuscany, Florence, Italy b
d
Received 18 July 2005; received in revised form 27 December 2005; accepted 3 January 2006 Available online 18 May 2006
Abstract Background: Clinical trials suggested feasibility and safety of early discharge after ST-segment elevation acute myocardial infarction (STEMI) for selected patients. Current United States and European guidelines recommend early discharge for uncomplicated AMI. The present study was aimed to assess early discharge in the current clinical practice for STEMI patients. Methods: Patients enrolled in the AMI-Florence Registry (Italy), a prospective, observational, population-based study performed in the early 2000s, were analysed. The proportion of STEMI patients eligible for early discharge and of those actually discharged early, patient features influencing early discharge and outcomes (mortality, reinfarction or urgent revascularization during the first year of follow-up) were assessed. Results: Of 751 STEMI patients, 59% were classified as eligible for early discharge, according to previously established criteria. Among patients considered eligible, those actually discharged early dropped from 33.1% in age < 70 years to 15.9% in age 70 – 79 years and to 11.7% in age 80+ years. Of eligible patients, 26% were actually discharged within 4 days. Age 70+ years (reference: 69 years or younger; 70 – 79 years: OR: 0.40, 95% CI: 0.22 – 0.73; 80+ years: OR: 0.33, 95% CI: 0.14 – 0.76) and diabetes (OR: 0.48, 95% CI: 0.24 – 0.98) were negative independent predictors; and coronary reperfusion (OR: 2.92, 95% CI: 1.54 – 5.53) or firstly admitted to teaching hospital (OR: 1.68, 95% CI: 1.03 – 2.74) were positive predictors, of the multivariate probability of being early discharged. Among patients eligible for early discharge, 1-year and 3-year mortality did not differ significantly between patients actually, and not, early discharged. After 1 year of followup, no deaths were observed among patients early discharged.
* Corresponding author. Tel.: +39 556263373; fax: +39 556263375. E-mail address: [email protected] (A. Barchielli). 1 AMI-Florence Working Group: Careggi Hospital: G.M. Santoro, N. Carrabba (Cardiology Unit 1), G. Santoro, G. Corti (Cardiology Unit 2), M. Margheri, S. Valente (Department of Critical Care Medicine and Surgery, Unit of Internal Medicine and Cardiology), F. Ferrante (Medicine Unit 1), V. Verdiani (Medicine Unit 2), I. Olivotto (Medicine Unit 3), N. Marchionni, M. Monami (Department of Critical Care Medicine and Surgery, Unit of Gerontology and Geriatric Medicine); Local Health Unit Florence: A. Barchielli, D. Balzi, P. Naldoni, C. Melani (Epidemiology Unit), L. Fratoni (Emergency Ambulance Transport 118 Service); Santa Maria Annunziata Hospital: A. Fantini (Coronary Care Unit), M. Torri (Medicine Unit 1), G. Regoli, C. Mugnaini (Medicine Unit 2); Santa Maria Nuova Hospital: M.C. Landini (Coronary Care Unit), M. Granelli (Medicine Unit 1); Nuovo San Giovanni di Dio Hospital: M. Filice (Coronary Care Unit), P. Fabiani (Medicine Unit 1); Mugello Hospital: F. Miglietta (Intensive Therapy Unit), L. Scarti, I. Berni (Medicine Unit); Serristori Hospital: G. Fabrizi de Biani, F.A. Tarmun (Medicine Unit); Regional Agency for Public Health of Tuscany: E. Buiatti, S. Arniani, S. Bartolacci (Epidemiology Unit). 0167-5273/$ - see front matter D 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijcard.2006.01.006
58
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
Conclusion: This study confirms, in a setting of current clinical practice, the reliability of criteria for identifying patients eligible for early discharge. Besides, the results confirmed the safety of this practice in selected patients. About 1/4 of eligible patients are actually early discharged, confirming the existence of important opportunities to improve the efficiency in STEMI management. D 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Acute myocardial infarction; Population-based registry; Prognosis; Early discharge
1. Introduction Improvement in medical care (e.g. recognition of the beneficial effects of early mobilization; increased use of coronary reperfusive therapy), and increasing pressure toward cost containment in health care, have resulted in a progressive shortening of hospital stay after acute myocardial infarction (AMI) [1]. In early 1990s, some clinical trials suggested the feasibility and the safety of early discharge after AMI (3– 4 days), at least for selected subgroups of patients [2– 4]. Simple clinical characteristics have been defined to identify cases of ‘‘uncomplicated infarction’’: these include absence of reinfarction, myocardial ischaemia, stroke, hearth failure, need for by-pass surgery, intra-aortic balloon pumping, emergency catheterization, cardioversion or defibrillation in the first four in-hospital days [5]. When applied to patients in the GUSTO-I trial, these criteria could identify patients at very low risk, who were considered eligible for early discharge from hospital [5]. In keeping with these criteria, current US and European guidelines recommend an early discharge for patients with uncomplicated AMI [6,7]. Recently, prognostic scores for ST-segment elevation myocardial infarction (STEMI) treated with primary angioplasty have been proposed [8,9]. In particular, the score proposed by De Luca et al. reliably identifies a large group of patients at very low risk, who may safely discharged early after primary angioplasty [8]. Cost/effectiveness analyses suggested that, in relation to conventional standards, prolonged hospital stay after thrombolysis or primary angioplasty of patients with uncomplicated myocardial infarction is ‘‘economically unattractive’’ [8,10] and an early discharge strategy for low-risk patients results in a substantial cost saving [2,11,12]. Despite this, Kaul et al. [13] recently have pointed out that, in the setting of some international clinical trials, no more than 40% of patients eligible for early discharge were actually discharged early. In particular, the rate of early discharge of eligible patients was especially low in the European Countries included in the analysis (Belgium, France, Germany, Spain and Poland). The Florence Acute Myocardial Infarction (AMI-Florence) Registry is a population-based, observational study carried out in the Florence health district (Tuscany region, central Italy) [14], enrolling STEMI. In this area, primary percutaneous coronary intervention (PCI) is the most widely preferred myocardial reperfusive treatment. The present
analysis of the AMI-Florence dataset was aimed at identifying in the ‘‘real world’’ patients eligible for early discharge on the basis of criteria previously specified [5] and to determine the proportion of those actually discharged early. Furthermore, this study analysed the clinical factors that influence early discharge in the current clinical practice and it verified outcomes in relation to the length of hospital stay.
2. Materials and methods 2.1. Registry design and setting The structure and setting of the AMI-Florence Registry have been detailed elsewhere [14]. Briefly, the Florence health district (about 800,000 inhabitants) comprises five community hospitals and one teaching hospital, the latter implementing high-volume program for primary PCI, operative 24 h/day, 7 days/week. Patients considered eligible for coronary reperfusion are treated with primary PCI or fibrinolysis. Patients admitted to the teaching hospital are treated with primary PCI, while those admitted to community hospitals, are usually transferred to the teaching hospital for primary PCI or, less frequently, receive fibrinolytic therapy onsite [14]. The AMI-Florence Registry was a prospective, observational, population-based study, which included all cases of AMI arriving alive to one out of six hospitals in the Florence health district, with no exclusion criteria (recruitment period: 2000 – 2001). Patients with a diagnosis of suspected AMI were prospectively screened for eligibility and enrolled in the study if AMI was confirmed by at least two of the following criteria: typical chest pain lasting > 30 min, rise in serum creatine phosphokinase activity at least twice above the normal upper limit within 72 h of symptom onset and evidence of AMI on the first ECG (ST-segment elevation 0.1 mV in two or more adjacent leads or new onset left bundlebranch block). Information on demographics, medical history, clinical and ECG features of AMI, time delay between onset of symptoms and hospital admission, treatment and outcome during hospitalization, was collected using standard case-report forms. In our country, at the time of patients enrolment, ethics committee approval and informed consent from patients were not required for observational studies.
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
59
surgery or died within 4 days from hospital admission, were excluded from the present analysis (169/920).
2.2. Patients categorization Using criteria similar to those previously established in the setting of clinical trials [5,13], STEMI cases enrolled in the AMI-Florence Registry (920 patients) were defined eligible for early discharge (within 4 days of admission) when none of the following events occurred during the hospital stay: reinfarction, myocardial ischaemia, stroke, shock, heart failure (Killip class > 1), by-pass surgery, intraaortic balloon pumping, emergency cardiac catheterization (for assessment of sign or symptoms of clinical instability), or need for cardioversion or defibrillation. Patients who had a history of stroke, those who recently had undergone major
2.3. Follow-up Major adverse events, including nonfatal reinfarction, urgent revascularization and death for all causes, were evaluated during the first year after the index AMI. To this purpose, subsequent hospitalizations were identified through the Regional Hospital Discharges database, recording all admissions to regional public and private hospitals. The original clinical records of re-hospitalized patients were checked to identify the occurrence of non-fatal reinfarction
Table 1 Patients eligible for early discharge: demographic and clinical characteristics by length of hospital stay Variable
Demographic Female gender Age (mean) Medical history Previous myocardial infarction Angina, onset >1 month Angina, onset 1 month CABG PCI Atrial fibrillation, paroxysmal Atrial fibrillation, chronic Conduction disorders Congestive heart failure Depression Peripheral artery diseases Renal failure, serum creatinine 1.5 – 2.5 mg/dl Renal failure, serum creatinine >2.5 mg/dl Chronic obstructive pulmonary disease Gastrointestinal diseases Cancer, onset 5 years Cancer, onset >5 years Anaemia Diabetes Hypertension Dyslipidaemia Smokers, current Smokers, former Hospital admission Directly admitted to hospital with PCI facilities Arrival modalities: referred by mobile care unit Admitted during office hours Time symptom/hospital arrival 6h Time hospital arrival/reperfusive therapy (median) Clinical presentation, AMI characteristics, treatment CPK peak 72 h (median) AMI anterior location, Q-waves AMI other location, Q-waves AMI any location, non-Q-waves Reperfusive therapy (primary PCI or fibrinolysis) Fibrinolysis Primary PCI Stenting rate (PCI patients) TIMI 3 grade flow after PCI
Length of hospital stay
p-value
4 days (115 cases)
5+ days (327 cases)
23.5% 60.5 years
29.1% 66.9 years
7.8% 13.9% 24.4% 0.0% 5.2% 0.9% 2.6% 1.7% 0.0% 3.5% 6.1% 0.9% 0.0% 6.1% 20.9% 0.9% 3.5% 0.9% 9.6% 41.7% 31.3% 51.3% 17.4%
14.7% 15.9% 18.4% 2.1% 4.3% 4.0% 2.8% 6.1% 1.2% 3.4% 8.3% 4.3% 2.1% 6.7% 19.9% 1.8% 4.0% 4.0% 20.0% 49.2% 30.0% 35.2% 20.2%
<0.008
66.1% 91.3% 67.8% 82.2% 30 min
43.4% 85.9% 61.5% 75.8% 42 min
<0.001 0.12 0.22 0.03 0.40
1280 U/l 25.2% 60.9% 13.9% 87.0% 0.0 87.0% 96.0% 96.0%
1255 U/l 28.8% 52.3% 19.0% 61.8% 6.4% 55.4% 95.0% 95.0%
PCI: percutaneous coronary intervention, CABG: coronary artery bypass graft, AMI: acute myocardial infarction, CK: creatine phosphokinase.
0.25 <0.001 0.059 0.37 0.11 0.68 0.26 0.063 0.23 0.95 0.45 0.059 0.81 0.82 0.75 0.10 0.012 0.17 0.79
0.82
<0.25 <0.001 <0.001 0.71 0.31
60
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
or need for urgent coronary revascularization. Reinfarction was diagnosed in the presence of the above reported criteria for the diagnosis of the index AMI. Coronary revascularization, performed with percutaneous intervention or surgery, was considered urgent when not performed on elective basis. Mortality was checked by consulting the Registry Office of the municipalities of residence. 2.4. Statistical analysis Differences across patients sub-groups were evaluated with the v 2 test for categorical variables and with the Student’s t-test for continuous variables. All tests were twosided and statistical significance was defined as a p < 0.05. Stepwise multivariate regression analyses (forward method, with p < 0.10 for entrance into and p > 0.15 for removal from the model) were performed to identify factors associated with early discharge and prognosis. Multivariate logistic regression was used to identify among the demographic and clinical characteristics of patients (reported in Table 1) the independent determinants of early discharge. The goodnessof-fit of the model was checked through the HosmerLemeshow test. Odds ratios (OR) and 95% confidence intervals (95% CI) were calculated. In addition, Cox regression was used to assess the relationship between early discharge and prognosis after adjustment for clinical and demographic variables with an independent prognostic effect. To this purpose, hazard ratios (HR) and 95% confidence intervals (95% CI) were calculated. The proportional hazard assumption was checked through the test proposed by Grambsch and Therneau, and the goodness-offit of the model through the test proposed by May and Hosmer [15]. Analyses were carried out using STATA statistical package (version 6.0, Stata Corporation, College Station, Texas, 1999).
3. Results 3.1. Patients characteristics Of 751 patients included in the present analysis (median age: 70 years, interquartile range: 60 –79), 442 (59%) were classified eligible for early discharge and 134 (17.8%) discharged alive within 4 days (Fig. 1). Because of the community composition of Florence health district, many patients enrolled in the Registry were elderly. In particular, cases considered not eligible were significantly older than those considered eligible for early discharge (mean age: 76.4 vs. 65.2 years, p < 0.001), the percentage of eligible decreasing from 75.8% in age < 70 years to 48.2% in age 70 –79 years and to 36.1% in age 80+ years. Besides, not eligible patients were more frequently females (35.6% vs. 27.6%, p = 0.020), affected by previous cardiovascular (i.e.: congestive heart failure: 7.0% vs. 0.3%, p < 0.001; angina of recent onset: 24.6% vs. 15.4%, p = 0.004; atrial fibrillation:
Fig. 1. Patients enrolled in AMI-Florence Registry and included in the analysis, by subgroups.
11.0% vs. 5.9%, p = 0.001; conduction disorder: 11.7% vs. 5.0%, p = 0.001; peripheral artery disease: 14.2% vs. 7.7%, p = 0.004) and non-cardiovascular diseases (i.e.: depression: 7.4% vs. 3.4%, p = 0.013; renal failure: 10.0% vs. 5.0%, p = 0.026; chronic obstructive pulmonary disease: 13.3% vs. 6.6%, p = 0.002; diabetes: 27% vs. 16%, p = 0.008), less frequently admitted directly to the teaching hospital with PCI facilities (33.3% vs. 49.3%, p < 0.001), with a higher proportion of anterior STEMI (44.0% vs. 27.8%, p < 0.001), less frequently treated with reperfusion (48.9% vs. 68.3%, p < 0.001) and, as expected, with a worse prognosis (i.e.: 1-year mortality: 24.0% vs. 4.3%, p < 0.001). Among patients considered eligible for early discharge, 115 (26.0%) were actually early discharged. Patients discharged early were significantly younger and less frequently affected by diabetes (Table 1). Besides, patients discharged early showed a lower prevalence (borderline significant) of previous myocardial infarction, conduction disorders and renal failure. On the other hand, more frequently, they were admitted directly to the hospital with PCI facilities and treated with reperfusive therapy. Interestingly, coronary reperfusion (with primary PCI being the most common strategy: 93.0%, 281/302) was significantly more frequent among cases directly admitted to the teaching hospital with PCI facilities than to community hospitals (88.1% vs. 49.1%, p < 0.001). Besides, patients directly admitted to the hospital with PCI facilities were significantly less affected by conduction disorder (2.8% vs. 7.1%, p = 0.034) and more frequently admitted within 6 h from symptom onset (85.8% vs. 71.0%, p = 0.001), and borderline significantly younger (mean age: 64 vs. 66 years, p = 0.071). Among patients considered eligible, those actually discharged early dropped from 33.1% (91/275) in age < 70 years to 15.9% (17/107) in age 70 –79 years and to 11.7% (7/60) in age 80+ years. A large part of patients eligible for early discharge arrived at hospital within 6 h of symptoms of onset (0 – 3 h: 56.6%, 4 – 6 h: 21.7%), the remaining 21.7% arriving 7 h or
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
61
later. Patients eligible and actually discharged early decreased from 28.8% when admitted within 3 h to 27.1% when admitted within 4 –6 h and to 17.7% when admitted 7 h or later from symptoms onset. 3.2. Length of hospital stay of patients eligible for early discharge By definition, the length of stay during the first hospital admission was significantly shorter in patients discharged early compared to those late discharged (respectively: 3.4 and 9.1 days), with a mean difference of about 6 days, largely explaining the difference observed in the total length of hospital stay during the first year after the index AMI (respectively: 9.4 and 16.6 days) (Table 2). Patients discharged early during the first hospitalization showed a shorter length of stay for the following hospital admissions during the 1-year follow-up, but this difference was less relevant (about 1.5 days; respectively: 6.0 and 7.4 days). The length of hospital stay (Fig. 2) differed significantly between patients firstly admitted to teaching and community hospitals (mean: 6.8 vs. 8.5 days, p = 0.002) and, more markedly, between patients treated and not treated with reperfusive treatment (mean: 7.0 vs. 9.0 days, p < 0.001). The difference across the hospitals of first admission was not significant when the comparison was stratified by treatment.
Fig. 2. Patients eligible for early discharge: length of hospital stay stratified by hospital of first admission and reperfusive treatment (TH: teaching hospital, CH: community hospitals, R+: reperfusive treatment, R : conservative treatment).
Among patients aged <70 years, the multivariate logistic regression analysis showed that the probability to be discharged early was borderline significantly lower for patients with diabetes (OR: 0.49, 95% CI: 0.22 – 1.10, p = 0.083) and significantly higher for those treated with reperfusion therapy (OR: 3.74, 95% CI: 1.84 – 7.60, p < 0.001). Among those aged 70+ years, the probability to be discharged early was significantly higher among those directly admitted to the teaching hospital (OR: 2.78, 95% CI: 1.09– 7.11, p = 0.033).
3.3. Factors influencing early discharge at the multivariate analysis
3.4. Outcomes
At the multivariate logistic regression analysis, the probability to be discharged early was significantly lower among patients aged 70 years and older (reference: 69 years or younger; 70 – 79 years: OR: 0.40, 95% CI: 0.22 –0.73, p = 0.003; 80 years or older: OR: 0.33, 95% CI: 0.14 –0.76, p = 0.003) or with diabetes (OR: 0.48, 95% CI: 0.24 –0.98, p = 0.043), and higher among those treated with reperfusion therapy (OR: 2.92, 95% CI: 1.54 – 5.53, p = 0.001) or directly admitted to the teaching hospital (OR: 1.68, 95% CI: 1.03– 2.74, p = 0.038).
After 1 year of follow-up, patients discharged early had not significantly different rates of reinfarction and of the cumulative incidence of major adverse events than those discharged late (Table 2). On the other hand, whereas urgent revascularization was performed more often among patients discharged early (difference borderline significant), mortality was significantly lower for patients discharged early. In the multivariate Cox analysis, the difference in urgent revascularization between the two groups of patients was not significant ( p = 0.14). No deaths in the discharged early
Table 2 Patients eligible for early discharge: days spent in hospital and outcomes by length of hospital stay Variable
All patients (442)
Length of hospital stay 4 days (115 cases)
Length of hospital stay (mean) First admission Other admissions in the first year of follow-up Total admissions in the first year of follow-up Outcomes (%) 1-year reinfarction 1-year urgent revascularisation 1-year mortality 1-year major adverse eventsa 3-year mortality a
Death, reinfarction or urgent revascularisation.
7.6 7.1 14.7 4.1 13.8 4.1 20.8 9.5
3.4 days 6.0 days 9.4 days 3.5% 19.1% 0.0% 22.6% 1.7%
p-value
5+ days (327 cases) 9.1 days 7.4 days 16.6 days 4.3% 11.9% 5.5% 20.2% 12.2%
<0.001 0.31 <0.001 0.71 0.054 0.010 0.48 0.001
62
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
group were observed in the first year of follow-up; therefore, the Cox analyses were only performed after 3 years of follow-up. Whereas in the univariate Cox analysis, patients discharged early showed a significant lower risk of death (HR: 0.13, 95% CI: 0.03 –0.56, p = 0.006), at the multivariate Cox analysis, after adjusting for the other variables with an independent prognostic effect (age, reperfusive therapy and medical history of anaemia, renal failure or depression), the difference in mortality was no more significant (HR: 0.42, 95% CI: 0.10– 1.82, p = 0.25). It is noteworthy that, among patients aged <70 years and discharged with 4 days, no deaths were observed after 3 years of follow-up. Only 1.7% (2/115) of the patients discharged early had an adverse event (reinfarction or urgent revascularization) after discharge and within the ninth day from the hospital admission (the mean number of days of hospital stay for patients discharged later).
4. Discussion The results of this population-based study carried out in the early 2000s showed that a large proportion (59%) of STEMI patients fulfilled criteria previously established by Newby et al. [5] for being candidates to early hospital discharge. In our study, the proportion of patients eligible for early discharge is similar to those reported in some clinical trials for other European countries, which ranged from 57% in France to 70% in Poland in the ASSENT-2 trial, carried out in the late 1990s [13]. However, in these countries, only 2% or less of eligible patients were actually discharged early. This proportion was very low, particularly when compared to the about 40% of early discharges that has been reported in USA and New Zealand [13]. In our study, about 26% of eligible patients were actually early discharged, a proportion greater than those previously reported in other European countries. Some issues must be discussed which are likely to account, at least in part, for this large difference. In Italy, where the National Health System covers hospital care without any charge for patients, a diagnosis-related-groups (DRGs) payment system for hospital reimbursement was introduced since the middle 1990s. As a consequence, substantial efforts have been made to improve performance and efficiency in hospital management, resulting in a general shortening in the length of hospital stay. Furthermore, our data was collected some years after the ASSENT-2 trial, when spreading of updated guidelines in the clinical practice might have contributed to reduce the length of hospital stay after uncomplicated AMI. Nevertheless, about 2/3 of eligible patients were discharged after 5 or more days in hospital suggesting that, as previously observed [13], important opportunities still remain to improve efficiency in AMI care. Our data were derived from an observational study that enrolled all patients with STEMI from the target population
who were hospitalized during the recruitment period, without selection bias. This is an important reason of interest for our results. Indeed, whereas previous analyses were referred to the selected setting of clinical trials [13], our study was aimed at analysing data collected in the perspective of current clinical practice and, hence, referring to the ‘‘real world’’. In particular, older patients are often underrepresented in clinical trials [16], whereas in the ‘‘real world’’ a large proportion of AMI patients are older [17]. For example, the median age is 61 years in the ASSENT-2 trial [13], compared to 70 years in study (where about 52% of patients were aged 70 years). In our series, patients eligible for early discharge, according to Newby et al. [5] criteria, dropped from 76.8% in age < 70 years to 36.1% in age 80+ years, and eligible patients actually discharged early from 33% (91/275) to 11.7% (7/60). The multivariate analysis showed that ‘‘eligible’’ patients had a significantly lower probability of being actually discharged early when aged 70 years or affected by diabetes. As expected in the clinical practice, some characteristics of AMI patients that are not effectively captured using standardized criteria may influence physicians’ final decision about hospital discharge. This finding suggests that the presence of noncardiac factors (such as diabetes or of other important comorbidities) should be considered to improve the reliability of criteria for identification of low-risk patients eligible for early discharge. Nevertheless, even after exclusion from analysis of cases with diabetes, the proportion of eligible patients who were actually discharged early moderately increased to 28%, reinforcing the view that the efficiency of management of AMI can still be improved. On the other hand, the probability of early discharge was higher in patients directly admitted to the hospital with PCI facilities or treated with reperfusion therapy. From a general point of view, both factors can influence early discharge practice. In fact, some hospital may have a more pressing interest in shortening the length of hospital stay, so that it can increase the number of hospitalizations. On the other hand, coronary reperfusion (and, in particular, primary PCI) is currently recognized as the most appropriate and beneficial treatment for AMI [6,7,18] and use of beneficial treatments is likely to reduce the length of hospital stay [19,20]. In particular, regardless of clinical risk status, the success of PCI has been proved to be a strong independent predictor of 30-day major adverse cardiac events, and an early discharge policy for patients with successful PCI and less than three high-risk features has been suggested [19]. In our dataset, admission hospital and reperfusive treatment have a certain degree of colinearity. In fact, about 88% of patients directly admitted to the teaching hospital were treated with reperfusive treatment compared to 49% in community hospitals. In a previous analysis of the whole series, the hospital of admission was the strongest independent predictor for subsequent PCI, the higher perception of indication to reperfusion by attending physician in hospital with PCI facilities likely explain this finding [14]. In the present
A. Barchielli et al. / International Journal of Cardiology 114 (2007) 57 – 63
analysis, the result of multivariate analysis suggested that both reperfusive therapy and type of hospital of first admission were independent predictors of early discharge. Interestingly, the proportion of patients eligible for early discharge and of those actually discharged early was higher among patients early presented, suggesting an influence of time between symptoms onset and hospital arrival (and therefore to proper AMI treatments) also on the length of hospital stay. Hypothesizing the feasibility to apply an early discharge policy to all eligible patients, about 748 days spent in hospital could be potentially saved in age <70 years, 428 in age 70– 79 years and 261 in age 80+ years (respectively 22%, 18% and 8% of the total number of days spent in hospital during the first hospitalization). Mortality rates observed in this study confirmed, also at the level of current clinical practice, the reliability and the safety of criteria that originally have been determined in the setting of clinical trials for identifying low-risk patients eligible for early discharge [5,13]. Indeed, 1-year mortality of eligible patients was very low (about 4% for all subjects), and 1-year mortality and other clinically relevant adverse events (such as reinfarction or need for urgent PCI) did not significantly differ, within the group of eligible patients, between those who were and those who were not actually discharged early. It is noteworthy that, in this group of patients, no deaths were observed among discharged early patients after 1 year of follow up and that, after 3 years of follow-up, only 2 (1.7%) patients aged > 70 years died, whereas no deaths were reported among patients aged 70 years or less. In conclusion, our data confirm, in the current clinical practice, the reliability of criteria for identification of lowrisk patients eligible for early discharge. Besides, the results confirm the safety of this practice in selected patients. Despite this evidence, only about 1/4 of patients eligible to early discharge are actually discharged within 4 days of admission, confirming the existence of important opportunities to improve efficiency in AMI care.
References [1] Spencer FA, Lessard D, Gore J, et al. Declining length of hospital stay for acute myocardial infarction and postdischarge outcomes. Arch Intern Med 2004;164:733 – 40. [2] Topol EJ, Burek K, O’Neill WW, et al. A randomised trial of hospital discharge three days after myocardial infarction in the era of reperfusion. N Engl J Med 1988;318:1083 – 8. [3] Mark DB, Sigmon K, Topol EJ, et al. Identification of acute myocardial infarction patients suitable for early hospital discharge after aggressive interventional therapy. Results from the Thrombolysis and Angioplasty in Acute Myocardial Infarction Registry. Circulation 1991;83:1186 – 93.
63
[4] Sanz G, Betriu A, Oller G, et al. Feasibility of early discharge after acute Q wave myocardial infarction in patients not receiving thrombolytic treatment. J Am Coll Cardiol 1993;22:1795 – 801. [5] Newby LK, Califf RM, Guerci A, et al. Early discharge in the thrombolytic era: an analysis of criteria for uncomplicated infarction from the Global Utilization of Strepokinase and t-PA for Occluded Coronary Arteries (GUSTO) trial. J Am Coll Cardiol 1996;27:625 – 32. [6] Ryan TJ, Antman EM, Brooks NH, et al. 1999 update: ACC/AHA guidelines for the management of patients with acute myocardial infarction: executive summary and recommendations. Circulation 1999;100:1016 – 30. [7] Van de Werf F, Ardissino D, Betriu A, et al. Management of acute myocardial infarction in patients presenting with St-segment elevation. Eur Heart J 2003;28:28 – 66. [8] De Luca G, Suryapranata H, van’t Hof AW, et al. Prognostic assessment of patients with acute myocardial infarction treated with primary angioplasty. Circulation 2004;109:2737 – 43. [9] Morrow DA, Antman EM, Charlesworth A, et al. TIMI risk score for ST-elevation myocardial infarction: a convenient, bedside, clinical core for risk assessment at presentation. Circulation 2000; 102:2031 – 7. [10] Newby LK, Eisenstein EL, Califf RM, et al. Cost effectiveness of early discharge after uncomplicated acute myocardial infarction. N Engl J Med 2000;342:749 – 55. [11] Grines CL, Marsalese DL, Brodie B, et al. Safety and costeffectiveness of early discharge after primary angioplasty in low risk patients with acute myocardial infarction. J Am Coll Cardiol 1998;31: 967 – 72. [12] Bartholomew BA, Harjai KJ, Grines CL, et al. Variation in hospital length of stay in patients with acute myocardial infarction undergoing primary angioplasty and the need to change the diagnostic-related groups system. Am J Cardiol 2003;93:830 – 3. [13] Kaul P, Newby LK, Fu Y, et al. International differences in evolution of early discharge after acute myocardial infarction. Lancet 2004;363: 511 – 7. [14] Buiatti E, Barchielli A, Marchionni N, et al. Determinants of treatment strategies and survival in acute myocardial infarction: a population based study in the Florence district, Italy: results of the Acute Myocardial Infarction Florence Registry (AMI-Florence). Eur Heart J 2003;24:1195 – 203. [15] May S, Hosmer D. A simplified method of calculating an overall goodness-of-fit test for the Cox proportional hazard model. Lifetime Data Anal 1998;4:109 – 20. [16] Lee PY, Alexander KP, Hammill BG, et al. Representation of elderly persons and women in published randomized trials of acute coronary syndromes. JAMA 2001;286:708 – 13. [17] Parmley WW. Do we practice geriatric cardiology? J Am Coll Cardiol 1997;29:217 – 8. [18] Aversano T, Aversano LT, Passamani E, et al. Thrombolytic therapy vs. primary percutaneous coronary intervention for myocardial infarction in patients presenting to hospitals without on-site cardiac surgery. JAMA 2002;287:1943 – 51. [19] Heggunje PS, Harjai KJ, Stone GW, et al. Procedural success versus clinical risk status in determining discharge of patients after primary angioplasty for acute myocardial infarction. J Am Coll Cardiol 2004; 44:1400 – 7. [20] Dirksen MT, Ronner E, Laarman GJ, et al. Early discharge is feasible following primary percutaneous coronary intervention with transradial stent implantation under platelet glycoprotein IIb/IIIa receptor blockade. Results of the AGGRASTENT trial. J Invasive Cardiol 2005;17: 512 – 7.