The impact of short hospital stay on prognosis after acute myocardial infarction: An analysis from the ACSIS database

Abstract Background Current evidence regarding the optimal length of hospital stay (LOS) following myocardial infarction (MI) is limited. This study aimed to examine LOS policy for MI patients and to assess the safety of early discharge. Methods A prospective observational study that included patients with STEMI and NSTEMI enrolled in the Acute Coronary Syndrome Israeli Survey (ACSIS) during the years 2000–2016. Patients were divided into three subgroups according to their LOS: <3 days (short‐LOS), 3–6 days (intermediate‐LOS) and >6 days (long‐LOS). We compared baseline characteristics, management strategies and clinical outcomes at 30 days and 1 year in these groups. Results Ten thousand four hundred and fifty eight patients were enrolled in the study. The LOS of MI patients gradually decreased over time. Short‐LOS and intermediate‐LOS patients had similar clinical characteristics while patients in the long‐LOS group were older with more co‐morbidity. There was no difference in the clinical outcomes, including re‐MI, arrhythmias, 30 days MACE, and 30 days mortality between the short‐LOS and intermediate‐LOS groups. However, the rate of re‐hospitalizations was higher in the short‐LOS group (20.9% vs. 17.8%, p = .004) without evidence of increased cardiovascular events. In multivariate analysis, the LOS did not predict either 30 days mortality (HR: 1.3; CI:0.45–5.48), nor MACE at 30 days (HR: 1.1; CI:0.79–1.56). Conclusion Our study suggests that an early discharge strategy of up to 3 days from admission is safe for low and intermediate‐risk patients after both STEMI and NSTEMI. Nevertheless, this strategy is associated with an increased risk of potential avoidable readmission and there might be psychological and social factors that may warrant a longer stay.


| INTRODUCTION
Until the 1950s, patients with myocardial infarction (MI) were treated by immobilization and prolonged hospital stay (4-6 weeks). [1][2][3] This policy gradually changed in the following decades with a median hospitalization of 21 days in 1970, 14 days in 1980, and less thereafter. 4 Improvement in the management of acute MI over the past several decades, in conjunction with a policy of early mobilization, has led to a dramatic decline in overall mortality with a steady decrease in the length of hospital stay (LOS). [5][6][7] Though decreased LOS provides a significant cost savings for health care systems, there is concern that this policy might put patients at risk due to premature discharge.
The available evidence regarding the optimal LOS after a MI is scarce. The guidelines of the European Society of Cardiology 8 currently recommend that the optimal LOS should be determined on an individual basis, according to the patient's cardiac risk, comorbidities, functional status, and social support.
Several studies have shown that patients with STEMI who underwent successful primary PCI and complete revascularization can be safely discharged from hospital within 48-72 h. 3,[9][10][11][12][13][14] These low-risk patients can be recognized using various risk score systems 9-13 such as the Second Primary Angioplasty in Myocardial Infarction (PAMI-II) criteria 9 or the Zwolle primary PCI Index. [10][11][12][13] Only a few studies which investigated the safety of early discharge 9,11,12,15,16 post MI were randomized and most of them were relatively small and underpowered. Moreover, there are no actual guidelines recommendations concerning the optimal LOS following NSTEMI. [17][18][19] In the present study, we used the large ACSIS (Acute Coronary Syndrome Israeli Survey) database, in order to assess whether an early discharge strategy is safe, and to define the characteristics of patients who could benefit the most from this strategy.

| Design
All patients were segregated according to their LOS (LOS): up to 3, 3-6, and over 6 days. The short LOS was chosen based on previous studies and guidelines of early discharge after STEMI. 8,9,11,12 The Additional comparisons were made between STEMI and NSTEMI patients and according to the GRACE-score (Global Registry of Acute Coronary Events Risk Score) tertiles.

| Statistical analysis
Differences between groups were tested with chi-square for categorical variables and with t-test or Mann-Whitney test as appropriate for normally or non-Gaussian distributed continuous variables. Logistic regression, presented as a forest plot, was calculated to assess the relationship between baseline characteristics and LOS to the risk of re-hospitalization.
Survival curve plots were conducted using Kaplan-Meier estimator and differences between survival curves were tested using the Log-rank test. In order to explore the effect of LOS on survival or on 30 days MACE, logistic or Cox models adjusted for selected covariates were utilized.
Statistical analyses were performed using R Core Team (2015). 20 Statistical significance was defined as p-value <.05.

| Baseline and clinical characteristics
Baseline characteristics of the three groups are shown in Table 1.
Patients in the long LOS group (>6 days) were older with more comorbidities and risk-factors than patients in the intermediate and short LOS groups. The two latter groups were comparable, with a similar GRACE risk score (Table 1 and Figure 1A).

| Outcomes
Major clinical outcomes in the short and intermediate LOS groups are shown in Table 2. The rate of re-hospitalization was higher in the short-LOS group. All other outcome endpoints including MACE and mortality were not different between groups. These results were similar when sub-divided into STEMI and NSTEMI patients (data not shown), or when sub-divided between early period (2000-2008) and late period (2010-2016) (Supplementary Table 2). Interestingly, rehospitalization was significantly higher only in the early period.
One-year mortality rate was also not different between groups ( Figure 1B, p = .36).
We analyzed clinical outcomes by dividing the study population into tertiles according to their GRACE risk score (see Supplementary   Table 3). As expected, the rate of adverse events and mortality increased with increasing GRACE score. However, no statisticallysignificant difference in outcomes was found between the short-LOS and intermediate-LOS groups except for the rate of re-hospitalization which was higher in the highest GRACE score (>115) group (p = .03).
We performed a multivariate analysis to identify significant predictors for re-hospitalization within 30 days. Short length of stay, diabetes mellitus and a reduced ejection fraction were independently significant predictors for re-hospitalization ( Figure 2). The vast majority of these patients had an uncomplicated clinical course and in many cases stayed in hospital for observation as part of conservative and traditional policies. Indeed, we found that both groups had comparable baseline characteristics (Table 1). Surprisingly in the short-LOS group there were more patients with risk factors and known ischemic heart disease. This may be due to a smaller need for evaluation time and medical adjustments for already medically treated patients or because patients with recurrent events are more familiar with the procedures and treatments and consequently do not require long hospital stay.
The short-LOS group consisted of patients who were in slightly better clinical condition upon admission with fewer in-hospital complications (which were low in both groups). More importantly, the clinical outcomes at 30 days as well as the 1-year mortality were not different between the groups. Nevertheless, the rate of re-hospitalizations was higher in the short-LOS group but without evidence of cardiovascular events. Same results of higher rate of re-hospitalizations after short LOS were noted in previous studies. 22 The risk for re-hospitalization correlated with the GRACE risk score and was very low for patients with GRACE score under 115. The rate of re-hospitalization decreased throughout the years of the study, and correlation with short-LOS became non-significant in the later years (2010-2016). Rehospitalization was more common in older patients and in patients with diabetes mellitus and a reduced ejection fraction. This may indicate that a more conservative approach for these sub-populations is warranted.
Our study supports current guidelines advocating discharge 48-72 h after admission for an uncomplicated STEMI. 8 However, there is inconsistency in the literature about the optimal score for assessing the risk after MI in order to determine the desired length of stay.
Most studies defined the threshold for early discharge as an extremely low risk in STEMI patients. 9 (Table 3).
Furthermore, there are no guideline recommendations for optimal LOS following NSTEMI [17][18][19] and our study supports an approach of early discharge after an uncomplicated course.

| Limitations
The major limitation of our observational study is that we cannot demonstrate causality but only an association between variables and outcome, along with the risk of overlooked confounders. In addition, the data was collected over a 16-year period during in which various changes in therapies and intervention policies occurred that may have influenced the changes in LOS and prognosis.
The data were collected by many investigators, which may lead to different interpretations of the study questionnaire. In addition, readmissions were not divided into emergent and elective, and therefore may include some elective admissions such as planned PCI to a non-culprit vessel. Finally, we could not account for patients admitted for very short time because of leaving against medical advice, a factor that may also portend worse outcomes secondary to noncompliance.

| CONCLUSIONS
Our study suggests that an early discharge strategy of up to 3 days from admission is safe for low and intermediate-risk STEMI and NSTEMI patients (GRACE score <115). The cause of increased rehospitalization in some populations remains unclear and might be related to non-medical conditions. Potentially, more patients may benefit from an early discharge policy and shorter hospital stay, a policy that may be particularly relevant during the COVID-19 pandemic.