© 2015, Wiley Periodicals, Inc. DOI: 10.1111/joic.12254
PERCUTANEOUS CORONARY INTERVENTION Percutaneous Coronary Interventions and Hemodynamic Support in the USA: A 5 Year Experience NILESHKUMAR J. PATEL, M.D., 1 VIKAS SINGH, M.D., 2 SAMIR V. PATEL, M.D., 3 CHIRAG SAVANI, M.D., 4 NILAY PATEL, M.D., 5 SIDAKPAL PANAICH, M.D., 6 SHILPKUMAR ARORA, M.D., 7 MAURICIO G. COHEN, M.D., 2 CINDY GRINES, M.D., 6 and APURVA O. BADHEKA, M.D., F.A.C.P., C.C.D.S. 8 From the 1Staten Island University Hospital, Staten Island, New York; 2University of Miami Miller School of Medicine, Miami, Florida; 3 Western Reserve Health Education, Youngstown, Ohio; 4New York Medical College, Valhalla, New York; 5Saint Peter’s University Hospital, New Brunswick, New Jersey; 6Detroit Medical Center, Detroit, Michigan; 7Mount Sinai St. Luke’s Roosevelt Hospital, New York, New York; and 8The Everett Clinic, Everett, Washington
Objectives: To compare the utilization and outcomes in patients who had percutaneous coronary interventions (PCIs) performed with intra-aortic balloon pump (IABP) versus percutaneous ventricular assist devices (PVADs) such as Impella and TandemHeart and identify a sub-group of patient population who may derive the most benefit from the use of PVADs over IABP. Background: Despite the lack of clear benefit, the use of PVADs has increased substantially in the last decade when compared to IABP. Methods: We performed a cross sectional study including using the Nationwide Inpatient Sample. Procedures performed with hemodynamic support were identified through appropriate ICD-9-CM codes. Results: We identified 18,094 PCIs performed with hemodynamic support. IABP was the most commonly utilized hemodynamic support device (93%, n ¼ 16, 803) whereas 6% (n ¼ 1069) were performed with PVADs and 1% (n ¼ 222) utilized both IABP and PVAD. Patients in the PVAD group were older in age and had greater burden of co-morbidities whereas IABP group had higher percentage of patients with cardiac arrest. On multivariable analysis, the use of PVAD was a significant predictor of reduced mortality (OR 0.55, 0.36–0.83, P ¼ 0.004). This was particularly evident in sub-group of patients without acute MI or cardiogenic shock. The propensity score matched analysis also showed a significantly lower mortality (9.9% vs 15.1%; OR 0.62, 0.55–0.71, P < 0.001) rate associated with PVADs when compared to IABP. Conclusion: This largest and the most contemporary study on the use of hemodynamic support demonstrates signi ficantly reduced mortality with PVADs when compared to IABP in patients undergoing PCI. The results are largely driven by the improved outcomes in non-AMI and non-cardiogenic shock patients. (J Interven Cardiol 2015;28:563–573)
Introduction Intra-aortic balloon pump (IABP) is the most commonly utilized mechanical support device in the
Nileshkumar J. Patel and Vikas Singh contributed equally to this manuscript. Disclosures: None. Funding source: None. Address for reprints: Apurva O. Badheka, M.D., F.A.C.P., C.C.D.S., Yale School of Medicine, 333 Cedars Street, New Haven, CT 06520, USA. Fax: þ1-203-737-2437; e-mail:
[email protected]
Vol. 28, No. 6, 2015
cardiac catheterization laboratories. The last decade has witnessed the development and Food and Drug Administration (FDA) approval of alternative percutaneous ventricular assist devices (PVADs) such as Impella (Abiomed, Danvers, MA) and TandemHeart (Cardiac Assist, Inc., Pittsburgh, PA). PVADs are capable of providing a more robust hemodynamic support when compared to IABP thus effectively improving cardiac index, mean arterial pressure and pulmonary capillary wedge pressure. As noted by the recently released 2015 SCAI/ACC/HFSA/STS clinical expert consensus statement on the use of mechanical
Journal of Interventional Cardiology
563
PATEL, ET AL.
support devices, the patient population which may benefit from these devices include those with acute decompensated systolic heart failure, cardiogenic shock as a bridge to recovery, those undergoing high-risk (unprotected left main, multivessel, severely depressed left ventricular function) percutaneous coronary interventions (PCI), those with cardiogenic shock after acute myocardial infarction (AMI).1 In relation to PCI, the current practice guidelines from the American College of Cardiology Foundation (ACCF) and the American Heart Association (AHA) provide a Class IIa recommendation for the use of IABP for patients with cardiogenic shock after AMI who do not quickly stabilize with pharmacological therapy, Class IIb for alternative PVADs in patients with refractory cardiogenic shock and Class IIb for elective insertion of an “appropriate” hemodynamic support device in carefully selected high-risk patients2,3 The benefit of IABP in these patients has been questioned in randomized clinical trials.4,5 The hemodynamic benefit of PVADs translating into decreased mortality when compared to IABP has also not been shown.6–10 Despite the lack of clear evidence, the utilization of PVAD has increased substantially in the last decade.11 The present study was designed to provide further insights into PCIs performed with hemodynamic support (IABP or PVAD) using the nation’s largest available hospitalization database. The goals of the study were to (1) compare the utilization and outcomes in patients who had PCIs performed with IABP versus PVADs and (2) identify a sub-group of patient population who may derive the most benefit from the use of PVADs over IABP.
Methods The Nationwide Inpatient Sample (NIS) is the largest available all-payer database of hospital inpatient stays in United States including data on approximately 7–8 million discharges per year. Each individual hospitalization is de-identified and maintained in the NIS as a unique entry with 1 primary discharge diagnosis and up to 24 secondary diagnoses during that hospitalization. Each entry also carries information on demographic details, insurance status, co-morbidities, primary and secondary procedures, and hospitalization outcome. The NIS contains the clinical and resource use information included in a typical discharge abstract, with safeguards to protect
564
the privacy of individual patients, physicians, and hospitals (as required by the data sources). Data from the NIS have been used to identify, track and analyze national trends in healthcare usage, patterns of major procedures, access, disparity of care, trends in hospitalizations, charges, quality, and outcomes.11–13 This was a cross sectional study using the NIS database between the years 2008–2012. Procedures were identified through appropriate clinical modification of International Classification of Diseases, ninth edition (ICD-9-CM) codes. We identified patients who had PCI (ICD9 procedure code: 36.07 and 36.06), PVAD (ICD9 procedure code: 37.68 and 37.62), and IABP placement (ICD9 procedure code: 37.61). Both Impella and TandemHeart devices were considered together in the PVAD category as they share the same ICD-9 code. We included patients who had PCI and underwent PVAD or IABP placement during the same hospital admission from 2008 through 2012. In order to restrict our evaluation to the use of PCI in a typical adult population, only participants >18 years of age were included. We excluded all observations with a missing data for age (n ¼ 15), sex (n ¼ 14), and mortality during hospitalization. This study involved de-identified data and was therefore exempted from institutional review board. We defined severity of co-morbid conditions using Deyo’s modification of Charlson’s co-morbidity index (CCI).14,15 This index contains 17 co-morbid conditions with differential weights. The score ranges from 0 to 33, with higher scores corresponding to greater burden of co-morbid diseases. Procedural complications were identified by Patient Safety Indicators (PSIs) which have been established by the Agency for Healthcare Research and Quality to monitor preventable adverse events during hospitalization. These indicators are based on ICD-9-CM codes and Medicare severity Diagnosis-Related Groups and each PSI has specific inclusion and exclusion criteria.16,17 PSI individual measure technical specifications, Version 4.4, March 2012 was used to identify and define preventable complications.18 In order to prevent classification of a pre-existing condition (e.g., stroke or heart block) as a complication, cases with the ICD-9-CM code for a complication listed as the principal diagnosis (DX1) were excluded. Annual hospital volume was determined using the unique hospital identification number to calculate the total number of PCIs performed by a particular institution in a given year. In order to assess impact of annual
Journal of Interventional Cardiology
Vol. 28, No. 6, 2015
HEMODYNAMIC SUPPORT IN PCI
hospital volume on outcomes and for the ease of interpretation, annual hospital volume was divided into quartiles. This methodology of identifying patients undergoing procedures, co-morbid conditions and associated complications has previously been utilized in several studies.12,13,19 Statistical Analysis. Stata IC 11.0 (Stata-Corp, College Station, TX) and SAS 9.3 (SAS Institute Inc, Cary, NC).19 All analyses were performed using hospital-level discharge weights provided by the NIS to minimize biases. Hierarchical mixed effects models were generated in order to identify the independent multivariate predictors of in-hospital mortality and complications. Three level hierarchical models (with patient level factors nested within hospital level factors) were created with the unique hospital identification number incorporated as random effects within the model. Subgroup analyses were also performed for PCIs with hemodynamic support in four different patient population (1) MI and cardiogenic shock. (2) cardiogenic shock only (without MI), (3) MI only (without cardiogemic shock), and (4) patients without MI or cardiogenic shock. We used propensity-scoring method to establish matched cohorts in order to control for imbalances of patients’ and hospitals’ characteristics between the 2 groups which may have influenced the primary outcome. A propensity score was assigned to each hospitalization and the following variables were adjusted for: age, gender, charlson’s comorbidity score, AMI, cardiogenic shock, teaching status of hospital, hospital bed-size, hospital volume in quartiles, and primary payer. Patients with similar propensity score in the 2 treatment groups were matched using a 1–3 scheme without replacement using greedy algorithm.
Results A total of 18,094 procedures were identified over the 5 year study period between 2008 through 2012. IABP was the most commonly utilized hemodynamic support device (93%, n ¼ 16, 803) whereas 6% (n ¼ 1,069) were performed with PVADs and 1% (n ¼ 222) utilized both IABP and PVAD. Table 1 demonstrates the baseline characteristics of the study population. There were significant differences between the patient population in IABP and PVAD groups. Patients in the PVAD group were older in age (64.3% were
Vol. 28, No. 6, 2015
>65 years old vs 50%) and had greater burden of co-morbidities including hypertension (68% vs 59%, P < 0.001), diabetes (45% vs 32%, P < 0.001), congestive heart failure (69% vs 44%, P < 0.001), obstructive pulmonary disease (22% vs 18%, P < 0.001), renal failure (30% vs 15%, P < 0.001) when compared to those in the IABP group. A higher proportion of patients in the PVAD group were admitted emergently (27% vs 11%, P < 0.001) whereas IABP group had higher percentage of patients with cardiac arrest (18% vs 9%, P < 0.001) and weekend admissions (25.5% vs 13.4%, P < 0.001). The percentage of patients with cardiac arrest was highest in the group which utilized both IABP and PVAD in the same hospitalization (27.5%). We observed an overall in-hospital mortality rate of 19.8% (20.1% for IABP, 12% with PVAD, and 41% in IABP þ PVAD group). Overall complications rate for this patient population was 35.5% (36% for IABP vs 26% for PVAD vs 52% for the IABP þ PVAD group). The use of IABP was associated with higher cardiac (9% vs 4%) and respiratory (19% vs 11%) complications whereas PVADs were associated with higher rate of vascular complications (8.6% vs 5.5%). (Table 2A and B). Trend of in-hospital mortality are shown in Figure 1. On multivariable analysis, the significant predictors of reduced mortality were the use of PVAD (OR, 95% confidence interval) (OR 0.55, 0.36–0.83, P ¼ 0.004), having a private insurance (OR0.83, 0.70–0.99, P ¼ 0.03), higher household income (OR 0.77, 0.64–0.93, P ¼ 0.007) and procedures performed at higher volume centers (4th quartile: OR 0.65, 0.51–0.83, P < 0.001). Conversely significant predictors of increased in-hospital mortality were the use of both IABP and PVAD (OR 2.20, 1.20–4, P ¼ 0.01), older age, female gender, presence of cardiogenic shock, absence of insurance, and procedures performed at non-teaching hospitals (Table 3). Patients who underwent PCIs with hemodynamic support were subdivided into in four categories: (1) AMI and cardiogenic shock (n ¼ 6,658 of which 105 were PVAD recipients, (2) cardiogenic shock without AMI (n ¼ 2,856 of which 141 were PVAD recipients), (3) AMI without cardiogenic shock (n ¼ 4,393 of which 44 were PVAD recipients), and (4) patients without AMI or cardiogenic shock (n ¼ 4,187 of which 779 were PVAD recipients). The use of PVADs was a significant predictor of reduced mortality (OR 0.16, 0.07–0.36, P < 0.0001, as well as complications rate (OR 0.45, 0.32–0.64, P < 0.001) when compared to
Journal of Interventional Cardiology
565
PATEL, ET AL. Table 1. Baseline Characteristics of the Study Population Variable
Total observation (%) Age in years (%) 18–34 35–49 50–64 65–79 > ¼ 80 Sex (%) Male Female Race (%) White Black Hispanic Others Missing Cardiac arrest (%) Ventricular fibrillation (%) Comorbidities (%) Deyo modification of Charlson score (%) 0 1 More than or equal to 2 Obesity Hypertension Diabetes mellitus Congestive heart failure History of chronic pulmonary disease Peripheral vascular disease Neurological disorder Renal failure Deficiency anemia Median household income category for patient’s zip code (%) 1. 0–25th percentile 2. 26–50th percentile 3. 51–75th percentile 4. 76–100th percentile Primary payer (%) Medicare Medicaid Private including HMOs & PPOs Other/self-pay Hospital characteristics Hospital bed size (%)
566
IABP
PVAD
IABP þ PVAD
Overall
16,803 (92.9%)
1,069 (5.9%)
222 (1.2%)
18,094
0.8 11.3 38.0 35.4 14.6
0.3 5.1 30.3 42.3 22.0
1.4 8.1 39.2 33.8 17.6
0.7 10.9 37.5 35.8 15.1
69.2 30.8
74.7 25.3
77.5 22.5
69.6 30.4
66.5 6.3 7.3 7.9 12.1 18.0 17.6
64.2 8.2 6.9 9.0 11.7 8.7 7.3
55.9 12.6 8.6 10.8 12.2 27.5 16.2
66.2 6.5 7.3 8.0 12.0 17.6 16.9
2.1 30.4 67.5 12.1 59.1 32.4 43.6 17.8 9.7 4.7 15.0 18.5
5.2 19.3 75.5 11.0 68.2 45.2 69.2 22.4 21.7 3.6 29.7 23.5
1.4 21.2 77.5 13.5 56.3 39.6 64.4 14.0 16.7 9.9 23.0 24.3
2.3 21.2 68.1 12.0 59.6 33.2 45.4 18.1 10.5 4.7 16.0 18.8
25.0 26.8 25.0 23.3
32.9 23.1 21.7 22.4
18.8 23.2 26.1 31.9
25.3 26.6 24.9 23.3
48.9 7.8 32.4 10.9
65.8 6.8 22.0 5.4
46.9 7.7 34.2 11.3
49.9 7.7 31.8 10.6
P Value
