The prevalence of 30‐day readmission after acute myocardial infarction: A systematic review and meta‐analysis

Abstract Objective The 30‐day readmission is associated with increased medical costs, which has become an important quality metric in several medical institutions. This current study is aimed at clarifying the prevalence, the underlying risk factors, and reasons of the 30‐day readmission after acute myocardial infarction (AMI). Methods PubMed, Cochrane Library, and EMBASE were systematically searched to identify eligible studies. Random‐effect models were employed to perform pooled analyses. Means and 95% confidence intervals (CIs) were used to estimate prevalence and reasons for 30‐day readmission. We also used Odds ratios (ORs) to explore the potential significant predictors of risk factors of 30‐day readmission after AMI. Potential publication bias was assessed using funnel plot and Begg'test. Results A total of 14 relevant studies were included in this systematic review and meta‐analysis. The pooled 30‐day readmission rate of AMI was 12% (95% CI 0.11‐0.14). Acute coronary syndrome (ACS), angina and acute ischemic heart disease, and heart failure (HF) were the principal cardiovascular reasons of 30‐day readmission. Meanwhile, non‐specific chest pain was regarded as the significant cause among non‐cardiovascular reasons. The common co‐morbidities kidney disease, HF and diabetes mellitus were significant risk factors for 30‐day readmission. No significant publication bias was found by funnel plot and statistical tests. Conclusions The 30‐day readmission rate of post‐AMI ranged from 11% to 14% and can be mainly attributed to cardiovascular and non‐cardiovascular events. The common co‐morbidities, such as kidney disease, HF, and diabetes mellitus were significant risk factors for 30‐day readmission.

diagnosed AMI would have unplanned readmission during 30 days of hospital discharge, which estimated direct costs of $1 billion of annual Medicare expenditures in the United States. 2 Statistically, nearly 20% of Medicare beneficiaries was readmitted within 30-day after AMI. 3 Therefore, reducing the rates of rehospitalization has attracted attention from policymakers and medical workers as a way improve the quality of care and reduce costs, payment incentives and Medicare hospital readmission penalties were created to reduce readmission rates. 4,5 Beginning in year 2013, the Hospital Readmission Reduction Program (HRRP) was carried out by the US Centers for Medicare& Medicaid Services (CMS) to improve financial incentives for decreasing readmission, which hospitals received penalizing according to higher-than-expected risk-standardized 30-day readmission rates for heart failure, myocardial infarction, and pneumonia. 5 Re-hospitalization is a frequent negative outcome for both hospitals and patients, and is an enormous economic burden to the Medicare beneficiaries and private payer. 3 There was a weak but significant correlation between the reduction of the 30-day readmission and 30-day hospital mortality after hospital discharge. 6 The re-admission rate of 30-days after myocardial infarction reduced from 20.5% to 15.8% from 2001 to 2003 to 2009 to 2011, but this trend slightly decreased after adjusting patient characteristics and treatment methods. 7 Predicting risk factors and reasons of 30-day readmission after AMI could help clinicians to actively identify patients with the highest possibility to benefit from intensity of the readmission intervention, so as to optimize limited medical resources allocation and implement beneficial and sustainable intervention. 8,9 However, many previous studies were performed in single-center study with a small sample size and had shown inconsistent results for readmissions after 30-days of AMI. Hence，it is necessary to further understand the prevalence, potential causes, and risk factors for readmission. Therefore, the purpose of our study was aimed at clarifying the prevalence, identify, and compare the potential risk factors and reasons for AMI of the 30-day readmission. Moreover, we discuss the potential intervention strategies to reduce the identified risk factors and causes of readmission.

| METHODS
This present systemic review was performed according to predesigned protocol, which was conducted under the Meta-analysis of Observational Studies in Epidemiology. 10

| Search strategy and study selection
Relevant literatures were systematically acquired from three electronic databases using the PubMed, EMBASE, and Cochrane Library (contained Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews, and the Database of Abstracts of Reviews of Effect) for study prevalence of 30-day readmission after AMI. The articles published date from inception to May 26, 2019 to obtain any possible inclusion. The two reviewers independently performed title words to search eligible articles that included following two concepts: (a) readmission (readmission*, re-admission*, rehospitalization*, re-hospitalization*, reattendance*, re-attendance*, readmittance*, re-admittance*), and (b) acute myocardial infarction   (myocardial infarction, MI, acute myocardial infarction, AMI, non-STsegment elevation myocardial infarction, NSTEMI, ST-elevation myo-cardial infarction, STEMI). A total of 1104 studies were identified. The detailed search strategies were provided in the Appendix S1.

| Study inclusion and exclusion criteria
This systematic review and meta-analysis of inclusion criteria was directed on the basis of the Preferred Reporting Items for Systematic Review and Meta-analysis protocols (PRISMA-P

| Data extraction and methodological quality assessment
The standard EXCEL forms were used to extracted relevant data including study period, diagnoses, country, data source, study population, demographic characteristics, simple size, prevalence, definition of 30 day readmission, underlying risk factors as well as causes of 30-day readmission. Using multivariable analysis of risk factors and reasons of readmission identified in two or more studies was collected. The quality of included studies was assessed using standard from Critical Appraisal of the Health Research Literature: Prevalence or Incidence of a Health Problem, 11 which including eight items (ie, sample size, sample design, sampling frame, study and setting, measures, unbiased assessors, response rate and refusers, and prevalence rates) and one point for each item. The quality of individual studies was classified low-quality when total score is less than 6 or highlyquality if whole score is in 6 or more scores.

| Statistical analysis
The primary outcome is the prevalence of 30-day readmission and the secondary outcomes include underlying causes and risk factors for 30-day readmission after AMI. Random-effect models were employed to perform pooled analyses because clinical heterogeneity across contained studies. Means and 95% confidence intervals (CIs) were used to estimate prevalence and reasons for 30-day readmission. We conducted subgroup analyses by stratifying of region, study population, quality of included studies find out the sources of heterogeneity. We also used odds ratios (ORs) to explore the potential significant predictors of risk factors of 30-day readmission after AMI. Potential publication bias was assessed using funnel plot and Begg'test. The two articles 12,13 of data source acquired from Nationwide Readmissions Database (NRD) which is a large national database. Using influence analysis was performed to explore whether the overlap study population would influence the overall pooled readmission rate. If the Pvalue is less than .05, we consider the correlation to be statistically significant. All meta-analysis was performed using State 12.

| Search results and study characteristics
The overall of 1104 articles were identified in this meta-analysis after systematically searching, which used PubMed, Cochrane Library, and with NSTEMI and other two assessed STEMI patients. The sample size of each included study was different. We divided the sample size into three layers according to sample size less than 1000, 1000-10 000, and more than 10 000. Six studies included all patients, six articles included adults (aged≥18 years), and only one article included middle-aged crowd (aged 18-64 years), or elderly patients (aged ≥65 years). Four articles of population were defined as unplanned readmission and other were defined as readmission which was no strict distinction between planned readmission and unplanned readmission in the original study. Details baseline characteristics of selected studies were presented in Table 1. Three studies were identified as low-quality and another 11 as high-quality (Appendix S2).

| Publication bias
No significant publication bias was found by funnel plot and statistical test (Begg test, P = .274; Appendix 3). However, asymmetric funnel plots suggested potential publication bias in the current metaanalysis.

| DISCUSSION
To the best of our knowledge, this study is the initial systematic review and meta-analysis of all reported assessed prevalence of  Notes: NS, no statistical significance; + AMI READMITS score (first-day model); COPD, chronic obstructive pulmonary disease; HF, heart failure; kidney disease included renal failure, renal function (Cr > 2 mg/dL),acute kidney injury, end state renal disease/hemodialysis; MI, myocardial infarction; blood system disease refers to iron deficiency anemia or other anemia other rather leukemia. *Statistical signifcance (P < .05).
fluctuate，so the population of the two overlapping studies did not affect the overall readmission rate. We also identified that kidney disease, female sex, diabetes mellitus, COPD，HF is the predictor of early readmission. Initially, chronic kidney disease is commonly correlation with dyslipidemia, diabetes, and hypertension which result in atherosclerosis and endothelial dysfunction, so it is considered as intensively independent risk factor of patients diagnosed AMI. 26 The second risk factor is gender. Indeed, past research has shown that women were at higher risk of post-AMI 30-day readmission than men, especially younger women. 27 Females probable have different pathophysiological and clinical characteristics from men, 19 for example，women often have experience chest pain or myocardial ischemia rather than coronary artery obstruction. 27 Then, the risk factors for AMI in patients with diabetes mellitus were more than twice as high as in patients without diabetic of AMI. 28 The AMI people with renal insufficiency and diabetes mellitus are relevant to major adverse cardiac events and risk of unfavorable prognosis, which can offer worthwhile information for early risk stratification. 29  readmission had poor discrimination with a median C-statistic of 0.65, and were of uncertain universality due to methodological quality limitations. 33 Additional review suggested that there is no effective model to measure the re-admission rate of hospital or to establish readmission risk model for individual patients. 34 Most studies which lacked of models validation or being a single-center study were evaluated low to moderate quality. In order to avoid the limitation of the individual, we would use the method of pooled multiple studies to provide consistently identified variables.
This review was to summarize the potential risk factors and reasons by useful available literature to generate strategies to reduce 30-day readmissions. Early outpatient physical follow-up has been reviewed as an effective strategy to prevent readmission. However, in American hospitals with higher early follow-up rates after AMI do not effectively reduce the 30-day admission rate，so it is necessary to adopt other targeted strategies besides early doctor follow-up to reduce the readmission rate of this population. 35 In particular, we should take into account research on predictors of readmission, risk stratification, interventions, risk-standardized model to reduce 30-day readmissions.

| LIMITATIONS
This review of results has certain several limitations. First, because of different articles have various classifications and grouping of causes and risk factors, there are inconsistent definitions for the studied variables, which make combining them for meta-analysis difficult. For example, different age groups were studied and used to make it impossible to merge and analyze data. Similarly, many risk factors and causes of readmission were unclearly defined, which potentially lead to overlap or deletion of data. Second, the substantial heterogeneity between studies may be come from age, data source, study period, diagnose, study population as well as study designs. Due to the limitations of the data, we could not conduct subgroup analysis for each variable. As mentioned above, the majority of studies did not offer complete data, which some studies only probed readmission rates and did not include analyses of risk factors and causes.

| CONCLUSIONS
In conclusion, the 30-day readmission rate post-AMI ranged from 11% to 14% and can mainly come from cardiovascular and noncardiovascular reasons. ACS, angina and ischemic heart disease, heart failure, and acute myocardial infarction were the principal cardiovascular reasons of 30-day readmission. Meanwhile, non-specific chest pain was regarded as the significant cause among non-cardiovascular reasons. The common comorbidity such as kidney disease, HF, COPD, and diabetes mellitus were significant risk factors for 30-day readmission. Therefore, our finding can help develop targeted strategies and prediction model to lower readmission rates according to the underlying risk factors and reasons of the 30-day readmission after AMI.