Etiologies and predictors of 30‐day readmissions in patients undergoing percutaneous mechanical circulatory support–assisted percutaneous coronary intervention in the United States: Insights from the Nationwide Readmissions Database

Abstract

Background

Patients undergoing percutaneous mechanical circulatory support (pMCS)‐assisted percutaneous coronary intervention (PCI) represent a high‐risk group vulnerable to complications and readmissions.

Hypothesis

Thirty‐day readmissions after pMCS‐assisted PCI are common among patients with comorbidities and account for a significant amount of healthcare spending.

Methods

Patients undergoing PCI and pMCS (Impella, TandemHeart, or intra‐aortic balloon pump) for any indication between January 1, 2012, and November 30, 2014, were selected from the Nationwide Readmissions Database. Patients were identified using appropriate ICD‐9‐CM codes. Clinical risk factors and complications were analyzed for association with 30‐day readmission.

Results

Our analysis included 29 247 patients, of which 4535 (15.5%) were readmitted within 30 days. On multivariate analysis, age ≥ 65 years, female sex, hypertension, diabetes, chronic lung disease, heart failure, prior implantable cardioverter‐defibrillator, liver disease, end‐stage renal disease, and length of stay ≥5 days during index hospitalization were independent predictors of 30‐day readmission. Cardiac etiologies accounted for ~60% of readmissions, of which systolic or diastolic heart failure (22%), stable coronary artery disease (11.1%), acute coronary syndromes (8.9%), and nonspecific chest pain (4.0%) were the most common causes. In noncardiac causes, sepsis/septic shock (4.6%), hypotension/syncope (3.2%), gastrointestinal bleed (3.1%), and acute kidney injury (2.6%) were among the most common causes of 30‐day readmissions. Mean length of stay and cost of readmissions was 4 days and $16 191, respectively.

Conclusions

Thirty‐day readmissions after pMCS‐assisted PCI are common and are predominantly associated with increased burden of comorbidities. Reducing readmissions for common cardiac etiologies could save substantial healthcare costs.

1. INTRODUCTION

The Centers for Medicare & Medicaid Services (CMS) has identified 30‐day readmission as an important quality metric. Readmission after percutaneous coronary intervention (PCI) is common and is associated with significant healthcare costs and adverse patient outcomes.1, 2, 3 Thirty‐day all‐cause readmission rates after PCI have ranged from 4.6% to 15.6% in previous studies.1, 2, 3, 4 However, patients undergoing PCI represent a heterogeneous group of patients ranging from low‐risk to high‐risk patients. Patients undergoing PCI and percutaneous mechanical circulatory support (pMCS) represent one such high‐risk group that is highly vulnerable to complications and readmissions. A recent analysis from the National Cardiovascular Data Registry showed an increasing trend of pMCS‐assisted PCI in recent years.5 Moreover, with evolving technologies and techniques focusing on complex coronary interventions, it is anticipated that such procedures will continue to increase.6 However, data on readmission rates and predictors of readmissions in this high‐risk cohort are lacking. Thus, we sought to determine the incidence, predictors, causes, and costs of 30‐day readmissions in patients undergoing pMCS‐assisted PCI in a contemporary, nationally representative patient population in the United States.

2. METHODS

2.1. Data source

The study cohort was derived from the Nationwide Readmission Database (NRD), a publicly available database of all‐payer hospital inpatient stays developed by the Agency for Healthcare Research and Quality as part of the Healthcare Cost and Utilization Project. We used the NRD databases from 2012 to 2014. The NRD was constructed from 22 states with reliable, verified patient linkage numbers in the State Inpatient Databases that could be used to track the patient across hospitals within a state, while adhering to strict privacy guidelines. This database includes approximately 14 million patients, accounting for 51.2% of the total US resident population and 49.3% of all US hospitalizations. National estimates are obtained using sampling weights provided. Patients have a unique identifier that allows each patient to be tracked (the variable named “NRD_visitlink”). We determine the time between the first admission and the readmission by using the variable “NRD_daystoevent” and calculating the difference between that variable and the length of stay. A detailed explanation of all the variables in the NRD is available online (https://www.hcup-us.ahrq.gov/nrdoverview.jsp).

This study was deemed exempt by the institutional review board, as the NRD is a publicly available database that contains de‐identified patient information.

2.2. Study cohort

We identified the study population using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD‐9‐CM) procedure codes for PCI with stent placement (36.06, 36.97) and either percutaneous ventricular assist device (pVAD; 37.68 and 37.62) or intra‐aortic balloon pump (IABP; 37.61) in the 2012–2014 NRD (n = 48 680 across 698 centers). We did not include patients who had extracorporeal membrane oxygenation, as it was not possible to ascertain if the patient underwent percutaneous vs surgical extracorporeal membrane oxygenation. We excluded records of patients that were admitted during the month of December (n = 4235), as there will be incomplete data for those patients who may be readmitted during January of the next year, as this database includes only patients admitted during a specific calendar year; patients age < 18 years (n = 2); patients who died during index admission (n = 9278; 24.1%); same‐admission‐day transfers (n = 5705); second or third readmission of a single patient (n = 197); and patients who did not have mortality data (n = 16; Figure 1).

Figure 1.

Flowchart of study population. Abbreviations: PCI, percutaneous coronary intervention; pMCS, percutaneous mechanical circulatory support

2.3. Patient and hospital characteristics

Baseline patient characteristics included were age, sex, primary expected payer, median household income, and relevant comorbidities (hypertension [HTN], diabetes mellitus [DM], obesity, congestive heart failure [CHF], end‐stage renal disease [ESRD], chronic pulmonary disease, peripheral vascular disease, hypothyroidism, depression, smoking, previous myocardial infarction, previous PCI, previous coronary artery bypass grafting, previous transient ischemic attack/cerebrovascular accident (TIA/CVA), known coronary artery disease [CAD], atrial fibrillation, previous permanent pacemaker, previous implantable cardioverter‐defibrillator (ICD), anemia, liver disease, fluid and electrolyte disorders, other neurological disorders, pulmonary disease, and cancer). The severity of comorbid conditions was defined using a validated Deyo modification of the Charlson Comorbidity Index (CCI).7, 8 Other characteristics such as teaching status of the hospital, median household income, insurance status, elective admission status, and discharge disposition were also included. Other variables extracted were in‐hospital characteristics such as length of stay, admission requiring intensive care unit, cardiogenic shock, TIA/CVA, cardiac arrest, acute kidney injury (AKI), AKI requiring dialysis, major bleeding, vascular and cardiac complications, and hospital costs. (For the ICD‐9‐CM codes used to identify and define these variables, see Supporting Information, Table 1, in the online version of this article.)

2.4. Study outcomes

Our primary outcome of interest was 30‐day all‐cause readmission. Both unplanned and planned readmissions were included. Complications were identified using Patient Safety Indicators and ICD‐9‐CM codes (see Supporting Information, Table 2, in the online version of this article). Vascular complications were defined using codes for accidental puncture, injury to blood vessels, creation of arteriovenous fistula, injury to retroperitoneum, vascular complications requiring surgery, hemorrhage or hematoma, and other vascular complications not elsewhere classified. Major bleeding was defined using codes for as gastrointestinal (GI) or other bleeding with shock requiring vasopressors or transfusion, hemopericardium requiring pericardiocentesis, contusion of abdomen or extremities with compartment syndrome, and central nervous system bleeding. The primary diagnosis of each readmission record was grouped into major diagnostic categories as specified by the CMS, which largely corresponds to single‐organ system or etiology9 (see Supporting Information, Table 3, in the online version of this article).

2.5. Statistical analysis

Univariate differences in baseline characteristics between readmitted and nonreadmitted patients were evaluated using χ² tests for categorical variables and 2‐sample t tests for continuous variables. To assess the predictors of readmission, multivariable logistic regression analyses were performed. Both univariate and multivariate predictors were analyzed. Variables with a P value <0.2 on univariate analysis were entered into a multivariable regression model with backward stepwise selection to identify independent predictors of 30‐day readmission. None of the variables, except total cost, had any missing data. Data on cost was missing for 322 (1.1%) index hospitalizations. Statistical analyses was performed with Stata software, version 13 (StataCorp LP, College Station, TX). A P value <0.05 was considered significant for all the analyses.

3. RESULTS

A total of 29 247 patients (based on the weighted sample) who had PCI and pMCS during index hospitalization and who survived were included in the analysis. Of the included patients, pVADs were used in 3846 patients, an IABP was used in 24 992 patients, and 409 patients had both pVAD and IABP. The mean age was 65 years, with 28% female. Baseline patient and hospital characteristics and selected in‐hospital outcomes are summarized in Table 1. About 86% of the patients presented with ACS, and almost half of the patients had cardiogenic shock. Compared with patients who were not readmitted within 30 days, those readmitted were slightly older, represented a higher proportion of females, and were more likely to have HTN, DM, CHF, chronic pulmonary disease, peripheral vascular disease, ESRD, atrial fibrillation, anemia, prior ICD, liver disease, fluid and electrolyte disorders, CCI score ≥ 3, and nonprivate insurance. Patients who were readmitted within 30 days had higher incidence of AKI and longer length of stay with higher total hospital costs, and they were more likely to be discharged to a nursing home/nursing facility. The incidence of major bleeding and vascular complications was similar in both groups.

Table 1.

Baseline characteristics and in‐hospital outcomes of patients undergoing pMCS‐assisted PCI who survived after index hospitalization^a

	Overall, N = 29 247	30‐Day Readmission		P Value
		Yes, n = 4535	No, n = 24 712
Age, y	65.3 ± 12.5	66.8 ± 12.4	65.0 ± 12.5	<0.001
Female sex	28.1	32.6	27.2	<0.001
Comorbidities
HTN	64.0	68.4	63.2	<0.001
DM	34.9	40.2	33.9	<0.001
ESRD	3.2	4.9	2.9	<0.001
CHF	50.4	56.7	49.3	<0.001
Obesity (BMI ≥30 kg/m2)	14.1	14.4	14.1	0.804
Chronic pulmonary disease	17.7	22.0	16.9	<0.001
PVD	12.3	15.4	11.7	<0.001
Hypothyroidism	8.4	9.2	8.3	0.225
Depression	6.2	6.9	6.0	0.243
Smoking	38.9	37.0	39.3	0.103
Prior MI	11.5	11.2	11.5	0.697
Prior PCI	13.6	13.2	13.7	0.631
Prior CABG	4.3	4.2	4.3	0.748
Prior CVA/TIA	5.8	6.7	5.6	0.049
Known CAD	77.8	79.0	77.5	0.230
Carotid artery disease	2.1	1.7	2.2	0.255
AF	19.2	22.1	18.7	0.001
Prior ICD	2.9	3.9	2.7	0.016
Prior PM	1.5	1.6	1.5	0.715
Anemia	19.5	23.3	18.8	<0.001
Liver disease	1.7	2.5	1.5	0.008
Cancer	2.4	2.7	2.4	0.420
Fluid and electrolyte disorders	33.3	36.9	32.6	0.002
Other neurological disorders	4.6	4.1	4.7	0.271
CCI score				<0.001
0	2.2	1.6	2.3
1	25.5	18.9	26.8
2	30.2	29.4	30.4
≥3	42.0	50.1	40.6
Other characteristics
Teaching hospital	61.6	59.7	61.9	0.115
Median HHI, US$				<0.001
1–39 999	26.9	31.9	25.9
40 000–50 999	26.5	26.3	27.7
51 000–65 999	24.7	21.5	25.2
≥66 000	22.0	18.9	22.5
Primary payer				<0.001
Medicare/Medicaid	61.4	59.6	71.1
Private insurance	28.0	20.4	29.4
Self‐pay/other	10.6	8.5	11.0
Discharge disposition				0.047
Home	80.3	78.7	80.6
Nursing home/skilled facility	15.9	17.8	15.6
Against medical advice	0.4	0.7	0.4
Transfer to another hospital	3.3	2.8	3.4
Clinical presentation
STEMI	60.9	59.2	61.3	0.04
NSTEMI	25.0	28.5	24.4	0.002
Elective admission	11.9	10.9	12.1	0.194
In‐hospital outcomes
LOS, d	6 (4–11)	7 (5–11)	6 (4–11)	<0.001
LOS ≥5 d	67.5	74.4	66.2	<0.001
ICU admission	24.7	25.0	24.7	0.812
Cardiogenic shock	48.8	50.3	48.5	0.229
Acute respiratory failure	27.9	29.8	27.5	0.099
Cardiac arrest	14.5	13.7	14.6	0.361
AKI	25.6	29.7	24.8	0.002
AKI requiring hemodialysis	1.7	1.7	1.7	0.881
Hospital costs	$38 723 (26 483–58 825)	$39 336 (27 583–59 641)	$38 606 (26 186–58 585)	0.132
Major bleeding	7.3	7.4	7.3	0.891
Vascular complications	2.7	2.8	2.7	0.761
Cardiac complications	7.9	6.6	8.2	0.058
CVA/TIA	2.1	2.3	2.1	0.734

Abbreviations: AF, atrial fibrillation; AKI, acute kidney injury; BMI, body mass index; CABG, coronary artery bypass grafting; CAD, coronary artery disease; CCI, Charlson Comorbidity Index; CHF, congestive heart failure; CVA/TIA, cerebrovascular accident/transient ischemic attack; DM, diabetes mellitus; ESRD, end‐stage renal disease; HHI, household income; HTN, hypertension; ICD, implantable cardioverter‐defibrillator; ICU, intensive care unit; IQR, interquartile range; LOS, length of stay; MI, myocardial infarction; NSTEMI, non–ST‐segment elevation myocardial infarction; PCI, percutaneous coronary intervention; pMCS, percutaneous mechanical circulatory support; PM, pacemaker; PVD, peripheral vascular disease; SD, standard deviation; STEMI, ST‐segment elevation myocardial infarction.

Data are presented as %, mean ± SD, or median (IQR).

Of the 29 247 patients who survived during index hospitalization, 4535 (15.5%) were readmitted within 30 days. The readmission rate was largely similar in patients who underwent pVAD‐assisted PCI vs patients with IABP‐assisted PCI (15.1% vs 15.6%; P = 0.665). The median days to readmission was 14 days (interquartile range, 8–22 days; see Supporting Information, Figure, in the online version of this article). There was a trend toward higher readmission rates in patients presenting with ACS vs elective admission (15.8% vs 14.0%; P = 0.08). Cardiovascular etiologies accounted for about 58% of readmissions (Figure 2). Among cardiovascular disease etiologies, systolic or diastolic heart failure (HF) accounted for 22% of readmissions, followed by stable CAD (11.1%), ACS (8.9%), and nonspecific chest pain (4.0%; Table 2). Of all 30‐day readmissions, 5.5% patients had concomitant cardiogenic shock. Noncardiac etiologies include sepsis/septic shock (4.6%), hypotension/syncope (3.2%), GI bleed (3.1%), AKI (2.6%), and pneumonia (2.3%). The univariate and multivariate predictors of readmission are shown in Table 3. On multivariate analysis, age ≥ 65 years, female sex, HTN, DM, chronic lung disease, HF, prior ICD, liver disease, ESRD, and length of stay ≥5 days during index hospitalization were independent predictors of 30‐day readmission in our cohort. Procedure‐related complications (major bleeding, vascular, or cardiac complications) were not associated with 30‐day readmissions (P > 0.10).

Figure 2.

Causes of 30‐day readmissions after pMCS‐assisted PCI according to system‐based major diagnostic categories. Abbreviations: PCI, percutaneous coronary intervention; pMCS, percutaneous mechanical circulatory support

Table 2.

Major etiologies of 30‐day readmission (among overall diagnoses)

Etiology	Incidence, %
CHF (systolic or diastolic)	22.0
Stable CAD	11.1
ACS	8.9
Sepsis/septic shock	4.6
Nonspecific chest pain	4.0
Hypotension/syncope	3.2
GI bleed	3.1
AKI	2.6
Pneumonia	2.3
Ventricular arrythmias	1.9
Acute respiratory failure	1.9
Complications of cardiac device	1.8
Othera	32.6

Abbreviations: ACS, acute coronary syndromes; AKI, acute kidney injury; CAD, coronary artery disease; CHF, congestive heart failure; GI, gastrointestinal.

^aEtiologies in aggregate, each accounted for ≤1.5%.

Table 3.

Univariate and multivariate predictors of 30‐day readmission in patients with PCI and pMCS

	Univariate Model			Multivariate Model
	OR	95% CI	P Value	OR	95% CI	P Value
Age ≥ 65 y	1.28	1.15‐1.44	<0.001	1.15	1.03‐1.3	0.017
Female sex	1.29	1.14‐1.47	<0.001	1.2	1.05‐1.37	0.009
HTN	1.26	1.11‐1.44	0.001	1.15	1.01‐1.31	0.04
DM	1.31	1.16‐1.47	<0.001	1.14	1.01‐1.3	0.04
HF	1.35	1.20‐1.52	<0.001	1.14	1.00‐1.29	0.05
Prior ICD	1.46	1.07‐1.98	0.016	1.43	1.03‐1.98	0.03
Chronic lung disease	1.29	1.21‐1.59	<0.001	1.25	1.08‐1.44	0.003
ESRD	1.74	1.33‐2.29	<0.001	1.42	1.07‐1.88	0.01
Liver disease	1.65	1.14‐2.39	0.008	1.57	1.09‐2.27	0.02
LOS >5 d	1.48	1.30‐1.68	<0.001	1.25	1.08‐1.45	0.002
Anemia	1.31	1.14‐1.51	<0.001	1.04	0.91‐1.20	0.58
PVD	1.37	1.17‐1.61	<0.001	1.16	0.99‐1.37	0.07
AKI	1.28	1.13‐1.45	<0.001	1.09	0.94‐1.27	0.24
AF	1.23	1.08‐1.41	0.002	1.1	0.96‐1.27	0.18
Electrolytes abnormalities	1.21	1.07‐1.36	0.002	1.03	0.91‐1.18	0.63

Abbreviations: AF, atrial fibrillation; AKI, acute kidney injuries; CI, confidence interval; DM, diabetes mellitus; ESRD, end‐stage renal disease; HF, heart failure; HTN, hypertension; ICD, implantable cardioverter‐defibrillator; LOS, length of stay; OR, odds ratio; PCI, percutaneous coronary intervention; pMCS, percutaneous mechanical circulatory support; PVD, peripheral vascular disease.

Among 4535 patients who had 30‐day readmission, 8.4% of patients underwent repeat PCI, of which 16% were elective PCIs. During readmission, pMCS was used in 1.1% of PCI procedures. One‐fifth of the patients developed AKI during readmission. The length of stay for readmissions was shorter than the index hospitalization, with a median of 4 days (interquartile range, 2–6 days). The total hospital cost for all 30‐day readmissions was $12 053 750 per year. The mean hospital cost per readmission was $16 191. Among the patients who underwent PCI during readmission, the readmission costs were higher, to $29 056.

4. DISCUSSION

We report a detailed analysis of 30‐day all‐cause readmissions after pMCS‐assisted PCI from a contemporary large nationwide database. The important findings of our paper are (1) about 15.5% of the patients undergoing pMCS‐assisted PCI were readmitted within 30 days; (2) age ≥ 65 years, female sex, hypertension, DM, chronic lung disease, HF, prior ICD, liver disease, ESRD, and length of stay ≥5 days during index hospitalization were independent predictors of 30‐day readmission; and (3) cardiac causes accounted for the majority of readmissions, with HF being the most common etiology, but noncardiac causes accounted for ~40% of readmissions.

Thirty‐day readmissions are considered a quality performance measure by the CMS. Previous studies based on registry data have reported 4.6% to 15.6% 30‐day readmission rates in patients undergoing PCI.1, 2, 3, 4 However, the prior studies included all‐comers populations, with only a small proportion of pMCS‐assisted PCI patients. In recent years, there has been a marked increase in the use of pMCS devices in patients undergoing PCI.5, 10 In our study of 29 247 patients who underwent pMCS‐assisted PCI from 2012 to 2014, the 30‐day all‐cause readmission rate was 15.5%. These rates are higher than the previous all‐comers PCI cohorts; however, this is not an unexpected finding, considering the high‐risk patient profiles requiring complex interventions. It is interesting to note that the readmission rate is similar to that in other high‐risk groups such as transcatheter aortic valve replacement11 and coronary artery bypass surgery.12, 13 Patients undergoing pMCS‐assisted PCI have a high burden of comorbidities, acute clinical presentation such as cardiogenic shock, and a high rate of periprocedural complications.10 Indeed, in our study, age ≥ 65 years, female sex, and comorbidities such as hypertension, DM, chronic lung disease, HF, prior ICD, liver disease, and ESRD were independently associated with 30‐day readmission. Length of stay ≥5 days was also a significant predictor for 30‐day readmission. This finding is likely because sicker patients require a longer stay and subsequently are more prone to readmissions. In our cohort, postprocedure vascular or bleeding complications, although relatively frequent, were not associated with readmissions. This finding needs to be interpreted in the context that major bleeding and vascular complications are associated with increased in‐hospital and 30‐day mortality14, 15, 16; hence, such patients get excluded from readmission statistics. Moreover, patients with vascular, bleeding, or cardiac complications would have required close monitoring and aggressive management, which may have played a role in preventing early readmissions. Another possible explanation could be just that our study lacked sufficient power to detect any difference. Future studies should focus on the implications of procedural complications on mortality, morbidity, and healthcare costs in this high‐risk cohort.

The majority of the patients in our cohort were admitted for acute myocardial infarction, with concomitant HF or cardiogenic shock presentation. This also reflects the real‐world practice in which IABP or pVAD (eg, Impella [Abiomed, Danvers, MA]) are increasingly used for urgent or emergent PCI.5, 10 A report from the National Cardiovascular Data Registry involving 76 474 patients who underwent PCI in the setting of cardiogenic shock between 2009 and 2013 found that pMCS was used in 46% of cases.17 On the other hand, the use of pMCS for elective PCI is less frequent and only limited to “high‐risk” PCI. The use of IABPs or pVADs in such cases is not supported by any evidence of clinical benefit or by professional guidelines, although several expert groups recommend its use in “high‐risk” PCI.18, 19

This is the first study to systematically explore the causes of 30‐day readmission after pMCS‐assisted PCI. The majority of the readmissions were due to cardiac causes (60%), with HF (22%) and stable CAD (11%) being the most common cause of all the readmissions. In previous studies of PCI, cardiac causes accounted for the majority of readmissions; however, the prevalence of HF readmissions was lower. In a study of 3 250 194 Medicare patients who underwent PCI, ischemic heart disease and HF were the most common causes of readmission, accounting for 25% and 15%, respectively.3 In another study of 9288 PCI patients, ischemic heart disease and HF accounted for 23% and 9% of the readmissions, respectively.2 In our cohort, more than half of our patients had a history of CHF and had cardiogenic shock at the time of presentation or during hospitalization, which may be in a part responsible for higher rates of HF readmissions. Early postdischarge follow‐up and other interventions focusing on reduction of HF and CAD may result in improvements in quality of care and reductions of costs within our healthcare system.20, 21 Noncardiac causes account for about 40% of the readmissions, with sepsis/septic shock, hypotension/syncope, and GI bleeding being the most common. High‐risk patients, need for multiple antithrombotic medications, and longer initial length of stay may be some of the plausible reasons responsible for these noncardiac readmissions. A multidisciplinary approach with equal emphasis on cardiac and noncardiac care is needed in these patients to ensure high‐quality care. Our results emphasize the need for further research to examine the preventability of 30‐day readmissions after pMCS‐assisted PCI, which represents one of the highest‐risk cohorts of patients undergoing PCI.

4.1. Study limitations

Our study should be interpreted in the context of several potential limitations. First, we used an administrative dataset, which lacks clinical and laboratory variables, angiographic and procedural characteristics, and medication information. Second, to capture concurrent use of pMCS during PCI, we only included patients who received pVAD or IABP on the same day of their PCI procedure. However, because data regarding the time of the placement of pMCS are not available in the NRD, it is possible that we included some patients who received pMCS after PCI (on the same day) either due to occurrence of procedural complications or due to operator preference. Moreover, we cannot distinguish Impella from TandemHeart (CardiacAssist Inc., Pittsburgh, PA) because both share the same ICD‐9 codes for pVAD. Third, the readmission rate might have been overestimated due to planned readmissions or coding errors.22 The causes of readmission were identified using the primary discharge diagnosis codes, and it was not possible to identify with certainty which readmissions were staged for subsequent PCI. Also, due to coding‐based nature of the database, we were unable to further analyze the admission diagnosis of CAD and chest pain. Fourth, we did not have data on long‐term outcomes beyond 30 days, which would be of interest in future studies. Lastly, and importantly, early mortality in this population is high and may have influenced the findings. Nonetheless, we included a contemporary nationwide database with a large sample size for a comprehensive analysis of etiologies and predictors of 30‐day readmissions after pMCS‐assisted PCI.

5. CONCLUSION

Thirty‐day readmissions after pMCS‐assisted PCI are common and are predominantly associated with increased burden of comorbidities. The cost of readmissions remains high, and measures should be undertaken to identify such high‐risk patients. Reducing readmissions for common cardiac etiologies could save substantial healthcare costs.

Author contributions

Chirag Bavishi, MD, MPH, and Alejandro Lemor, MD, contributed equally to this article.

Conflicts of interest

The authors declare no potential conflicts of interest.

Supporting information

eFigure 1. 30‐day readmission trend based on days to readmission

eTable 1. Variables by ICD 9 codes

eTable 2. Complications by ICD 9 code

eTable 3. ICD‐9 codes in groups for readmission etiologies

Bavishi C, Lemor A, Trivedi V, et al. Etiologies and predictors of 30‐day readmissions in patients undergoing percutaneous mechanical circulatory support–assisted percutaneous coronary intervention in the United States: Insights from the Nationwide Readmissions Database. Clin Cardiol. 2018;41:450–457. 10.1002/clc.22893