Author + information
- Received October 24, 2013
- Revision received June 13, 2014
- Accepted July 25, 2014
- Published online October 7, 2014.
- ∗Bonde Artificial Heart Laboratory, Yale University School of Medicine, New Haven, Connecticut
- †Section of Cardiac Surgery, Yale University School of Medicine, New Haven, Connecticut
- ‡Center for Advanced Heart Failure and Transplantation, Yale University School of Medicine, New Haven, Connecticut
- ↵∗Reprint requests and correspondence:
Dr. Pramod Bonde, Section of Cardiac Surgery, Center for Advanced Heart Failure and Transplantation, Yale University School of Medicine, 330 Cedar Street, Boardman 204, New Haven, Connecticut 06510.
Background The number of alternatives to intra-aortic balloon counterpulsation in the treatment of anticipated and established acute circulatory failure is growing. Despite the clinical importance and significant cost of short-term mechanical circulatory support (MCS) devices, the state of their present use has not been analyzed on a national scale.
Objectives The purpose of this study was to characterize the demographics, treatment practices, survival rates, and cost of short-term MCS.
Methods In this serial cross-sectional study, we analyzed all adult patients receiving short-term MCS in the United States from 2004 to 2011 by using the Nationwide Inpatient Sample from the Healthcare Cost and Utilization Project.
Results From 2007 to 2011, use of percutaneous devices for short-term MCS increased by 1,511% compared with a 101% increase in nonpercutaneous devices. Mortality rates declined over this period (p for trend = 0.027) from 41.1% in 2004 to 2007 to 33.4% in 2008 to 2011. A similar trend was observed for the subset of patients with cardiogenic shock, decreasing from 51.6% to 43.1% (p for trend = 0.012). Hospital costs also declined over this period (p for trend = 0.011). Multivariable analysis revealed balloon pumps (odds ratio [OR]: 2.00; 95% confidence interval [CI]: 1.58 to 2.52), coagulopathy (OR: 2.35; 95% CI: 1.88 to 2.94), and cardiopulmonary resuscitation (OR: 3.50; 95% CI: 2.20 to 5.57) before short-term MCS were among the most significant predictors of mortality.
Conclusions Use of short-term MCS in the United States has increased rapidly, whereas rates of in-hospital mortality have decreased. These changes have taken place in the context of declining hospital costs associated with short-term MCS.
Acute circulatory collapse is a broad term referring to failure of the pumping mechanism of the heart and an inability to maintain adequate organ perfusion. The most common situation in which it is encountered is cardiogenic shock. However, similar circulatory collapse can be anticipated during procedures that may compromise hemodynamic stability, including high-risk percutaneous coronary intervention (PCI), ablation for arrhythmias, and transcatheter valvular interventions.
Historically, institution of short-term mechanical circulatory support (MCS) was largely reserved for patients exhibiting significant circulatory compromise requiring cardiac output augmentation, with a bias toward supporting patients perceived as eligible for transplant or left ventricular assist device (LVAD) implantation (either bridge or destination therapy). More recently, the availability of rapidly deployable percutaneous MCS has led to a paradigm shift in the field that is characterized by growing use of these devices in an anticipatory or prophylactic fashion that was previously uncommon. Although the intra-aortic balloon pump (IABP) is not a true circulatory support device because it does not contribute directly to cardiac output, it was the only rapidly deployable support device available for decades. Reliable circulatory devices including the Thoratec PVAD (Thoratec Corporation, Pleasanton, California), AB5000 and BVS 5000 (both Abiomed, Inc., Danvers, Massachusetts), and various centrifugal pumps usually require a median sternotomy. Venoarterial extracorporeal membrane oxygenation (ECMO) delivered through peripheral cannulation with a centrifugal pump driver is another option but has always been reserved for emergency near-arrest situations. The morbidity and mortality associated with these devices are significant and restricted their use to potential transplant candidates.
The Impella 2.5 (Abiomed, Inc., Danvers, Massachusetts) and TandemHeart (CardiacAssist, Inc., Pittsburgh, Pennsylvania) are percutaneous MCS devices that can be deployed in the catheterization laboratory. The CentriMag (Thoratec Corporation, Pleasanton, California) and the Impella 5.0/CP have greatly increased ease of cannulation and device placement. These developments, among others, have theoretically made possible early deployment of short-term MCS before the downward spiral and inflammatory cascade associated with circulatory collapse can develop. Irrespective of surgical or percutaneous deployment, all temporary MCS devices share similar functional characteristics in terms of augmentation of cardiac output in liters per minute.
In U.S. and European guidelines for acute heart failure, the mainstays of therapy remain intravascular volume control, inotropes, and IABP (1,2). The combination of revascularization, antithrombotic therapy, and intensive care management has only modestly affected the mortality of cardiogenic shock in the last decade (3,4). Despite a paucity of randomized trials, recent guidelines have recommended the use of short-term MCS for profound hemodynamic compromise (1,5); these recommendations reflect the growing impact of short-term MCS on clinical practice and the lack of viable alternatives.
Unlike the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) and Extracorporeal Life Support Organization (ELSO) registries, which track outcomes and adverse events in patients receiving long-term MCS and ECMO, no such registry exists for short-term MCS. To begin addressing the existing deficit of information on patients receiving short-term MCS, we examined national trends in utilization.
The Nationwide Inpatient Sample, Healthcare Cost and Utilization Project (HCUP), under the auspices of the Agency for Healthcare Research and Quality (6), is the largest database of all-payer inpatient hospital stays in the United States. It approximates a 20% stratified sample of all nonfederal hospitals. All discharges from sampled hospitals are included, thus enabling the generation of national estimates. This study was deemed exempt by Yale University’s Institutional Review Board.
We included all adults ≥18 years old who were receiving short-term MCS between 2004 and 2011. Short-term MCS was defined by the International Classification of Diseases-ninth revision-Clinical Modification (ICD-9-CM) codes for percutaneous (37.68) or nonpercutaneous (37.60, 37.62, and 37.65) MCS in any procedure position (Online Figure 1). Nonpercutaneous devices included the Thoratec PVAD, AB5000, BVS 5000, and CentriMag. Percutaneous devices included the TandemHeart and Impella devices. IABP (37.61), ECMO (39.65), and percutaneous cardiopulmonary support (PCPS) (39.66) were excluded from our definition of short-term MCS. Permanent devices (37.52 and 37.66) included the HeartMate XVE and HeartMate II (Thoratec Corporation). Earlier years were not analyzed because ICD-9-CM codes do not distinguish short-term from permanent MCS devices before 2004.
Elixhauser comorbidities were generated from ICD-9-CM diagnosis codes using the HCUP Comorbidity Software (7). The sum of comorbidities for each record was reclassified as 0, 1, 2, or ≥3. Comorbidities present in ≥5% of all patients were reported.
We defined the indication for a hospital stay as the diagnosis listed in the primary position and categorized each using HCUP Clinical Classification Software (CCS). Level 3 CCS diagnoses constituting ≥5% of all indications for hospital stays were classified as discrete categories, whereas remaining diagnoses were merged into a single category of other miscellaneous indications. Mortality rates were defined as the percentage of patients who died before discharge.
We used HCUP cost-to-charge ratios for each hospital, based on information from the Centers for Medicare and Medicaid Services, to adjust the total charges for each patient and to estimate costs. When a hospital-specific cost-to-charge ratio was not available, we used the corresponding state-level ratio. These estimates were then adjusted for inflation by using the Consumer Price Index Inpatient Hospital Services inflation multiplier (8), with 2011 as the base year. Hospital costs incurred by permanent device implantation or heart transplantation were included in all calculations.
An abrupt change in utilization pattern was observed in 2008. Based on this finding, weighted means for demographic and hospital course characteristics were calculated and presented as illustrative data for 2004 to 2007 and for 2008 to 2011 in aggregate. To test for linear and curvilinear trends in disease characteristics and outcomes over time, we adopted a method of variance-weighted regression (9,10). This methodology incorporates the standard errors associated with the estimates of each year but does not assume homogeneity of variance. The threshold for including yearly estimates was a relative standard error <30%.
Data from 2005 were excluded from the mortality trend analysis because 2 of 53 hospitals caused a significant downward distortion of overall mortality in a manner inconsistent with all other years examined. Trend analyses for hospital costs and length of stay were performed before and after removing the top 1% of values, with similar results (latter data not shown).
Multivariable logistic and linear regression was used to calculate the association among independent variables, in-hospital mortality, and the total cost of hospital stay. Hierarchical models were used to account for clustering of cases by hospital. Length of stay and total cost of hospital stay were both log-transformed to achieve less positively skewed distributions. Elixhauser comorbidities affecting ≥5% of patients were included as independent variables if they were significantly associated with outcomes on bivariate analysis. The significance level was set a priori at a p value ≤0.05. Statistical analyses were performed using SPSS version 21.0 (IBM Corporation, Armonk, New York) (Online Methods).
An estimated 11,887 patients received short-term MCS between 2004 and 2011. Whereas use of temporary nonpercutaneous devices increased 101% from 2007 to 2011, percutaneous device use increased by 1,511%. Indeed, percutaneous devices exhibited the fastest growth of all forms of MCS in this period (Figure 1, Online Table 2).
Most patient-related and hospital characteristics remained similar over time (Table 1). Approximately two-thirds of patients were male and white. In 2008 to 2011, approximately one-half of patients were ≥65 years of age. Large urban teaching hospitals were the predominant setting.
Hospital course and disease characteristics
Acute myocardial infarction (AMI) was the most common primary diagnosis (Table 2, Online Table 3), and this predominance increased linearly over time (p for trend = 0.003), whereas other miscellaneous diagnoses (Online Table 4) made up a declining proportion of patients (p for trend = 0.024). Approximately one-half of all patients carried a diagnosis of cardiogenic shock. There was an increase in the proportion of patients with ≥3 Elixhauser comorbidities over time (p for trend = 0.001).
Length of stay and disposition
The mean duration of stay for congestive heart failure (CHF) decreased substantially over time (p for trend = 0.002) (Table 3). Trends in length of stay for AMI and coronary artery disease (CAD) did not reach statistical significance. The proportion of patients discharged home increased over time (p for trend = 0.001). This increase was accompanied by a concomitant decline in transfers to other facilities (p for trend = 0.019). In 2008 to 2011, patients with AMI, CHF, or other miscellaneous diagnoses were discharged home in approximately 40% of cases, compared with 70.7% of those with CAD (Table 4). Patients with CAD also had the lowest incidence of cardiogenic shock of all groups.
As illustrated in Figure 2, mortality rates of short-term MCS recipients have decreased over time (p for trend = 0.027) (Table 5). This trend was also observed in the subset of patients with cardiogenic shock (p for trend = 0.012). Detailed outcome data for this subgroup are presented in Online Table 1. Patients 65 to 79 years of age had the highest odds of mortality. Comorbidities conveying the greatest risk were coagulopathies and fluid and electrolyte disorders. Other predictors of death included a diagnosis of cardiogenic shock and use of IABP or cardiopulmonary resuscitation before short-term MCS. After adjusting for all other variables, later calendar years remained predictive of lower mortality.
The mean cost of hospital stays declined from 2004 to 2011 (p for trend = 0.011) (Table 6). Stratifying by diagnosis, costs declined over time for all groups except patients with heart valve disorders and other miscellaneous diagnoses. A patient’s age was found to have a strong inverse correlation with hospital costs. IABP use and in-hospital death were associated with cost increases of 25.2% (p < 0.001) and 28.5% (p < 0.001), respectively.
These national data demonstrate that use of short-term MCS increased rapidly between 2007 and 2011, accompanied by declining rates of in-hospital mortality. Concomitant reductions in hospital costs were observed during this period.
Who receives short-term mechanical circulatory support?
Recipients of short-term MCS have historically been patients with cardiogenic shock or circulatory collapse who were considered eligible for transplant or LVAD. Accordingly, delivery of short-term MCS most often took place in transplant or LVAD centers and included surgically implanted temporary devices. Providers were primarily cardiac surgeons experienced in transplant and LVAD procedures. Operations to implant these devices are usually associated with significant morbidity; the mortality of these patients has remained similar over time.
The availability of percutaneous devices profoundly changed the field of short-term MCS. The results of this study illustrate a paradigm shift characterized by increased use of short-term MCS in an anticipatory or prophylactic fashion (e.g., during high-risk PCI and ablation procedures carrying a likelihood of impaired cardiac output), as opposed to after the onset of circulatory compromise (Central Illustration). The most striking demonstration of this is the large number of patients with a primary diagnosis of CAD who receive short-term MCS. This change in the use of short-term MCS, particularly the evolution of a “nonshock” group of patients, has had a profound influence on aggregate outcomes.
Efforts to reduce the mortality rate of cardiogenic shock have had limited success. The SHOCK (Should We Emergently Revascularize Occluded Coronaries for Cardiogenic Shock) trial demonstrated that early revascularization in cardiogenic shock secondary to AMI improves 6- and 12-month survival but does not reduce 30-day mortality (11). The IABP-SHOCK II (Intra-aortic Balloon Pump in Cardiogenic Shock II) trial revealed no effect of IABP on 30-day or 12-month survival (3,12). Indeed, cardiogenic shock in AMI continues to carry an estimated mortality rate of at least 40% to 50% (13,14). Those patients receiving short-term MCS are among the most severely ill.
We identified high mortality in short-term MCS recipients with AMI or valvular heart disease but low mortality among patients with CAD. This difference is consistent with their respective prevalence of cardiogenic shock. The growing number of patients with CAD undergoing percutaneous MCS-assisted PCI resulted in a decline in overall MCS mortality over time.
Early identification of impending circulatory collapse coupled with rapid implementation of MCS may contribute to mortality reductions (15). Newer MCS devices have reduced the time between decision and implementation compared with earlier pulsatile devices that required expertise restricted to centers with heart transplant or LVAD programs. Guidelines need to be developed to help enable early identification of patients for whom IABP is likely to be insufficient.
The observed growth in MCS use likely results from a reduced treatment threshold. The shift from predominantly surgical to percutaneous device delivery has minimized procedural invasiveness and enhanced ease of delivery. It also has expanded availability because devices can be implanted in the catheterization laboratory without the need for surgical consultation. Ongoing technologic advances continue to increase the absolute additional output (l/min) achievable and help to obviate escalation to surgically implanted devices.
Quality of life and cost
The proportion of patients discharged home, without need for home health care, increased over time. Furthermore, the fraction of patients receiving permanent heart-assist devices during their hospital stay increased. From 2008 to 2011, 24.8% of the patients with a primary diagnosis of CHF received permanent devices, which have been shown to confer both a survival benefit and an improved quality of life (16).
These improved outcomes have been achieved in the context of declining hospital costs. Shorter durations of stay are likely a major factor driving this decline. The reduced proportion of patients discharged to other facilities suggests these shorter stays reflect more rapid recoveries rather than increased use of rehabilitation facilities. Implantation of permanent devices and heart transplants are associated with higher hospital costs and lengths of stay.
A somewhat surprising finding of our study was the 25.2% increase in the cost of hospital stay associated with IABP use. One possible explanation is that IABP use is associated with a delay in instituting more aggressive forms of circulatory support in patients who require it. Such delays could escalate the severity of cardiogenic shock and ultimately increase the intensity and cost of the care required by those patients.
Limitations of percutaneous short-term mechanical circulatory support devices
The technology of short-term MCS continues to evolve. Some devices may be insufficient to augment cardiac output, whereas others need specialized skills rare among even skilled interventional cardiologists. The amount of myocardial support required by patients in cardiogenic shock is subject to rapid fluctuation. However, the ability to escalate the degree of cardiac assistance provided by current-generation percutaneous devices remains limited. Prompt and repeated assessments to identify cannula positions, malfunctions, and changes in flow as a result of volume status are paramount to patients’ survival and reduction of adverse events.
Perhaps the most crucial aspects of management are assessments for signs of recovery and for subsequent withdrawal of short-term MCS. Understanding that extended support may be needed before adequate recovery can occur is critical to management and is a major limitation of prior randomized trials (17–19).
This study has several limitations inherent to a retrospective analysis. First, the use of ICD-9-CM codes to identify short-term MCS procedures did not allow an analysis of outcomes for individual devices. In addition, a substantial proportion of the study population (21.6%) was assigned a miscellaneous indication for hospital stay.
In terms of mortality, it is difficult to differentiate and estimate the impact of short-term MCS evolution from other concurrent changes during the study period, including efforts to reduce door-to-balloon times and length of inpatient stays. Length of stay and home healthcare requirements were higher in patients with CHF, although the higher rate of permanent device implantation in these patients may be a confounder. Finally, our analysis was limited to hospital costs. Increased use of permanent devices will confer downstream costs to the healthcare system, whereas reduced need for home health care and intermediate facilities at discharge will achieve cost savings.
Short-term MCS therapy has witnessed rapid growth since 2007, with concomitant reductions in both mortality rates and hospital costs. Systematic and longitudinal collection of data pertaining to short-term MCS therapy is warranted to ensure improved outcomes in the future.
COMPETENCY IN MEDICAL KNOWLEDGE: Percutaneous heart-assist devices are increasingly used to prevent circulatory compromise in patients at risk of acute hemodynamic decompensation.
TRANSLATIONAL OUTLOOK: Future advances in percutaneous mechanical circulatory assist technology should address the need for bedside deployment and reduced levels of hemolysis that currently limit the duration of support.
The authors would like to thank their statistician Elsa Su, BS, for ensuring the accuracy of this analysis.
For the statistical methodology as well as supplemental tables and a figure, please see the online version of this article.
This work was supported by the William W. Glenn Research Foundation, New Haven, Connecticut. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute myocardial infarction
- coronary artery disease and other heart disease
- Clinical Classification Software
- congestive heart failure
- extracorporeal membrane oxygenation
- Healthcare Cost and Utilization Project
- heart valve disorder
- intra-aortic balloon pump
- left ventricular assist device
- mechanical circulatory support
- percutaneous coronary intervention
- percutaneous cardiopulmonary support
- Received October 24, 2013.
- Revision received June 13, 2014.
- Accepted July 25, 2014.
- American College of Cardiology Foundation
- Yancy C.W.,
- Jessup M.,
- Bozkurt B.,
- et al.
- McMurray J.J.V.,
- Adamopoulos S.,
- Anker S.D.,
- et al.
- de Waha S.,
- Desch S.,
- Eitel I.,
- et al.
- Peura J.L.,
- Colvin-Adams M.,
- Francis G.S.,
- et al.
- ↵Healthcare Cost and Utilization Project (HCUP). HCUP Nationwide Inpatient Sample (NIS). 2004–2011. Agency for Healthcare Research and Quality, Rockville, MD. Available at: http://www.hcup-us.ahrq.gov/nisoverview.jsp. Accessed July 26, 2014.
- ↵Healthcare Cost and Utilization Project (HCUP). HCUP Comorbidity Software. 2004–2011. Agency for Healthcare Research and Quality, Rockville, MD. Available at: http://www.hcup-us.ahrq.gov/toolssoftware/comorbidity/comorbidity.jsp. Accessed July 26, 2014.
- ↵United States Bureau of Labor. Consumer Price Index (CPI). Available at: http://www.bls.gov/cpi/#data. Accessed July 26, 2014.
- ↵Healthcare Cost and Utilization Project (HCUP). Using the HCUP Nationwide Inpatient Sample to Estimate Trends. 2011. Agency for Healthcare Research and Quality, Rockville, MD. Available at: http://www.hcup-us.ahrq.gov/reports/methods/2006_05_NISTrendsReport_1988-2004.pdf. Accessed July 26, 2014.
- Goldberg R.J.,
- Spencer F.A.,
- Gore J.M.,
- Lessard D.,
- Yarzebski J.
- Cheng J.M.,
- Corstiaan A.,
- Hoeks S.E.,
- et al.
- Dixon S.R.,
- Henriques J.P.S.,
- Mauri L.,
- et al.
- Seyfarth M.,
- Sibbing D.,
- Bauer I.,
- et al.
- Thiele H.,
- Sick P.,
- Boudriot E.,
- et al.