Abstract
Background:
As the technology of ventricular assist devices continues to improve, the morbidity and mortality for patients with a ventricular assist device is expected to approach that of orthotopic heart transplantation. The present study was performed to compare perioperative outcomes, readmission, and resource utilization between ventricular assist device implantation and orthotopic heart transplantation, using a national cohort.
Methods:
Patients who underwent either orthotopic heart transplantation or ventricular assist device implantation from 2010 to 2014 in the National Readmission Database were selected.
Results:
Of the 12,111 patients identified during the study period, 5,440 (45%) received orthotopic heart transplantation, while 6,671 (55%) received ventricular assist devices. Readmissions occurred frequently after ventricular assist device implantation and orthotopic heart transplantation, with greater rates at 30 days (29% versus 24%, P = .005) and 6 months (62% versus 46%, P < .001) for the ventricular assist device cohort. Cost of readmission was greater among ventricular assist device patients at 30 days ($29,115 versus $21,586, P = .0002) and 6 months ($34,878 versus $20,144, P = .0106).
Conclusion:
Readmission rates and costs for patients with a ventricular assist device remain greater than their orthotopic heart transplantation counterparts. Given the projected increases in ventricular assist device utilization and limited transplant donor pool, further emphasis on cost containment and decreased readmissions for patients undergoing a ventricular assist device is essential to the viability of such therapy in the era of value-based health care delivery.
Introduction
Nearly 6.5 million Americans live with heart failure (HF), a condition accounting for an estimated $35 billion of annual health care expenditure in the United States.1 Mortality after inpatient admission for HF has been estimated to be as great as 35% within 1 year and 75% within 5 years.2 HF leads federal funding mandates, and a disproportionate amount of resources are aimed at the management of advanced HF. With the combination of an aging population and increasing burden of ischemic heart disease, the prevalence of end-stage HF continues to rise.3,4
Although orthotopic heart transplantation (OHT) is widely accepted as the gold standard therapy for end-stage HF, ventricular assist devices (VADs) have improved outcomes for patients with advanced HF in the past decade and have been used increasingly as bridge-to-transplantation (BTT) and destination therapies (DT).1, 5–8 Seco et al.9 demonstrated equipoise in survival, acute rejection, or allograft vasculopathy in their meta-analysis of short- and long-term outcomes between OHT and BTT therapies. No further differences were demonstrated in postoperative mortality, stroke, renal failure, or bleeding.
Although the implantation of VAD is considered safe and effective, adverse events during VAD support can lead to poor outcomes and multiple readmissions, a costly consequence for the patient and the health care system alike. As experience with using BTT and DT as a VAD, it is possible that durable VAD therapy could afford patients similar outcomes compared with OHT, thereby decreasing the dependence on the transplant donor pool.10 The present study was performed to compare resource utilization, mortality, and readmissions between patients receiving VAD and OHT, using a national cohort from 2010 to 2014.
Methods
Datasource
The National Readmissions Database (NRD) is a nationally representative, all-payer inpatient administrative registry of acute care hospitals in the United States, provided by the Healthcare Cost and Utilization Project in sponsorship with the Agency for Healthcare Research and Quality. It contains more than 17 million discharges with appropriate hospital weights to estimate more than 36 million annual US hospitalizations from 2010 to 2014. Patient-level diagnostic and procedural data, hospital characteristics, and estimates of inpatient hospital supercharges were derived from the database. Additional estimates of hospital cost-to-charge ratios and diagnosis-related group (DRG) adjustments were utilized to estimate hospitalization costs and account for disease severity. This study was deemed exempt by the Institutional Review Board of the University of California, Los Angeles.
Study population
Adult patients undergoing isolated OHT or VAD placement between January through June annually from 2010 to 2014 were sampled from the NRD. Study cohorts were identified using the International Classification of Diseases, 9th edition, clinical modification (ICD-9 CM) procedural codes for OHT (37.51) and VAD (37.66). Patient and hospital identifiers were randomized within each year. Thus, data for 6-month readmission risk was calculated based on patients undergoing primary surgery during the first 6 months of each year of data in order to allow for uniform and adequate follow-up. Patients undergoing concomitant mitral valve surgery and coronary artery bypass graft were excluded. Comorbidities and complications associated with cardiovascular disease and cardiac surgeries were identified using previously validated ICD-9 CM procedure codes.
Study outcomes
The primary study outcomes of interest were inpatient mortality and 30-day readmission. Secondary outcomes included duration of stay, overall cost of hospitalization, and postoperative complications, including stroke, myocardial infarction, infection, and arrhythmia. The NRD provides hospital charges for each admission, which are often several times greater than the actual costs of care because of the complex nature of reimbursement. Thus, the cost was estimated for each patient, using hospital-specific charge-to-cost ratios provided by the Agency for Healthcare Research and Quality from the Centers for the NRD. These estimates were further adjusted for through the use of the Healthcare Cost and Utilization Project (HCUP) indices of the DRG to account for variance in severity of hospitalization.
Statistical analysis
Cost, duration of stay, mortality, and postoperative complications were estimated, using hierarchical multivariable regression controlling for patient demographics, comorbidities, and hospital characteristics. Patient-level demographic characteristics included age, race, insurance type, income, and comorbidity evaluated using the Elixhauser Index.11 Additional comorbidities included angina; prior stroke; chronic renal, pulmonary, and liver disease; obesity, cardiogenic shock, coagulopathy, and frailty defined by ICD-9 codes.10,12 Bed size, teaching, and geographic location (urban versus rural) were included to adjust for hospital variability in OHT and VAD performance.5,13 Hierarchic regression adjusting for hospital covariance in the nested sampling design was utilized as recommended for the NRD database. Mortality, duration of stay, and log-transformed costs were modeled using logistic, Poisson, and linear distributions, respectively. Statistical analyses were performed using Stata 15 (StataCorp LP, College Station, TX).
Results
Patient demographics and clinical characteristics
During the study period, 12,111 patients were identified: 5,440 (45%) patients underwent OHT and 6,671 (55%) patients VAD implantation (Table I). Trend analysis demonstrated OHT rates to have increased marginally during the study, and rates of VAD implantation nearly doubled, surpassing OHT for the first time in 2011. VAD patients were more likely to be older (56 versus 52 years, P < .018), have a greater Elixhauser Index score (6.7 versus 5.7, P = .041), and be insured by Medicare. VAD patients were more likely to have comorbidities including chronic heart failure (CHF), cardiogenic shock, endocarditis, prior coronary artery bypass grafting (CABG), chronic kidney disease, chronic lung disease, and peripheral vascular disease. Patient income characteristics were not different among cohorts.
Table I.
Demographic characteristics of patients undergoing OHT versus VAD in NRD 2010–2014.
| OHT, n (%) | VAD, n (%) | P value | ||
|---|---|---|---|---|
| Discharges | 5,440 | 6,671 | ||
| Sex | ||||
| Male | 4,049 (74) | 5,214 (78) | .049 | |
| Female | 1,391 (26) | 1,456 (22) | ||
| Age | 51.9 (±6.45) | 55.5 (±6.60) | .018 | |
| Mean Elixhauser Index | 5.74 (±1.08) | 6.68 (±1.11) | .041 | |
| Payer | ||||
| Medicare | 1,966 (36) | 3,141 (47) | <. 001 | |
| Medicaid | 609 (11) | 690 (10) | ||
| Private insurance | 2,613 (48) | 2,559 (39) | ||
| Self-pay | 20 (0) | 54 (1) | ||
| No charge | 2 (0) | 2 (0) | ||
| Other | 182 (3) | 177 (3) | ||
| Median household income | ||||
| Lowest (0–25) | 1,352 (25) | 1,778 (27) | .167 | |
| Middle Low (26–50) | 1,358 (25) | 1,744 (27) | ||
| Middle High (51–75) | 1,267 (24) | 1,567 (24) | ||
| Highest (76–100) | 1,371 (26) | 1,459 (22) | ||
| Hospital classification | ||||
| Government | 714 (13) | 704 (11) | .105 | |
| Nonprofit | 4,713 (87) | 5,912 (89) | ||
| Private | 13 (0) | 55 (1) | ||
| Bed size | ||||
| Small | 156 (3) | 104 (2) | <. 001 | |
| Medium | 210 (4) | 393 (6) | ||
| Large | 5,075 (93) | 6,173 (93) | ||
| Comorbidities | ||||
| Prior stroke | 174 (3.2) | 287 (4.3) | .138 | |
| Hyperlipidemia | 1,854 (34.1) | 2,286 (34.3) | .931 | |
| Angina | 113 (2.1) | 90 (1.3) | .188 | |
| Coronary artery disease | 2,087 (38.4) | 2,629 (39.4) | .609 | |
| Cardiogenic shock | 1,687 (31) | 3,448 (51.7) | <. 001 | |
| Endocarditis | 535 (9.8) | 1,259 (18.9) | <. 001 | |
| Prior CABG | 74 (1.4) | 154 (2.3) | .035 | |
| Chronic lung disease | 447 (8.2) | 903 (13.5) | .012 | |
| Peripheral vascular disease | 215 (4) | 357 (5.4) | .014 | |
| Chronic kidney disease | 2,000 (37) | 3,023 (45) | <. 001 | |
| Chronic liver disease | 101 (1.9) | 102 (1.5) | .515 | |
| Diabetes | 306 (5.6) | 426 (6.4) | .379 | |
| Anemia | 3,106 (57.1) | 3,843 (57.6) | .839 | |
| Coagulopathy | 2,211 (40.6) | 2,198 (33) | <. 001 | |
| Frailty | 144 (2.6) | 310 (4.6) | .005 | |
| Obesity | 547 (10.1) | 1,071 (16.1) | <. 001 |
OHT, orthotopic heart transplantation; VAD, ventricular assist device; CABG, coronary artery bypass grafting.
Outcomes and resource utilization
Overall, patients receiving a VAD had an increased duration of stay (36.3 days versus 35.2 days, P = < 0.001) (Table II). Inpatient mortality in patients receiving a VAD nearly doubled that of OHT patients (10% versus 5.2%, P < .001). Patients receiving a VAD also had slightly increased costs of index hospitalization ($213,667 versus $177,128, P = .05). Readmissions occurred frequently after VAD implantation and OHT, with greater rates at 30 days (29% versus 24%, P = .005) and 6 months (62% versus 46%, P < .001) for the VAD cohort.
Table II.
Trends in main outcomes (unadjusted) in National Readmission Database 2010–2014.
| 2010 | 2011 | 2012 | 2013 | 2014 | P value | ||
|---|---|---|---|---|---|---|---|
| Index admission cost | OHT (n = 5,440) | $140,851 | $175,844 | $166,997 | $192,737 | $206,793 | .024 |
| VAD (n = 6,671) | $211,611 | $210,819 | $224,835 | $215,978 | $206,117 | .005 | |
| Duration of stay, mean | OHT (n = 5,440) | 31.1 | 31.6 | 31.6 | 40.1 | 41.7 | .012 |
| VAD (n = 6,671) | 38.2 | 36.1 | 37.2 | 36.1 | 34.8 | .002 | |
| Index admission mortality | OHT (n = 5,440) | 6.2% | 4.5% | 4.3% | 3.9% | 7.5% | .4741 |
| VAD (n = 6,671) | 10.5% | 8.2% | 12.4% | 10.1% | 9.3% | .4647 | |
| Readmission, 30 days | OHT (n = 5,440) | 27.5% | 25.4% | 23.9% | 22.9% | 21.3% | .4354 |
| VAD (n = 6,671) | 23.2% | 34.7% | 27.5% | 30.0% | 27.4% | .092 | |
| Readmission, 6 months | OHT (n = 5,440) | 52.2% | 46.0% | 44.9% | 44.9% | 42.2% | .3161 |
| VAD (n = 6,671) | 58.6% | 66.5% | 60.6% | 60.3% | 63.4% | .4884 |
OHT, orthotopic heart transplantation; VAD, ventricular assist device.
In subgroup analysis comparing patients with OHT versus OHT after BTT, the two groups had similar index admission mortality and readmission rates (Table III). Patients with OHT bridged with VAD had lesser durations of stay (26.5 days versus 40 days, P < .001) and costs of the index admission ($160,117 versus $186,709, P = .0321). Among patients undergoing OHT from January to June in 2014, the year in which day of operation became available in the NRD, there was no difference in the average postoperative duration of stay between patients with OHT versus OHT after BTT (22.4 days versus 22.9 days, P = .859).
Table III.
Main outcomes (unadjusted) in National Readmission Database 2010–2014.
| Heart replacement modality | Subgroup analysis | |||||
|---|---|---|---|---|---|---|
| VAD (n = 6,671) | OHT (n = 5,440) | P value | OHT only (n = 3638) | OHT after BTT (n = 1,802) | P value | |
| Index admission cost | $213,667 | $177,128 | .050 | $186,709 | $160,117 | .032 |
| Duration of stay, mean | 36.3d | 35.2d | <. 001 | 40.0 | 26.5 | <.001 |
| Index admission mortality | 10% | 5.20% | <. 001 | 5.7% | 4.4% | .399 |
| Readmission, 30 days | 29% | 24% | .005 | 24% | 24% | .893 |
| Readmission, 6 months | 62% | 46% | <. 001 | 46% | 45% | .689 |
OHT = orthotopic heart transplantation, VAD = ventricular assist device, BTT = bridge-to-transplant
Patients who underwent VAD implantation had a 17% lesser comorbidity-adjusted duration of stay than those having OHT (IRR = 0.83, CI 0.82–0.84, P < .001) (Table IV). Compared to patients receiving Medicare, hospital stay was less for patients having private insurance (IRR = 0.96, CI 0.94–0.7, P < .001) and those with self-pay (IRR = 0.89, CI 0.84–0.95, p < 0.001). History of stroke, congestive heart failure, cardiogenic shock, peripheral vascular disease, chronic liver disease, diabetes, and frailty were associated with increased durations of hospital stay.
Table IV.
Duration of stay adjusted for demographics, comorbidities and disease severity.
| Patient-level covariates | IRR | P value | |
|---|---|---|---|
| OHT | ref | ||
| VAD | 0.83 (0.82–0.84) | <. 001 | |
| Previous VAD | 0.8 (0.79–0.81) | <. 001 | |
| Female | 0.97 (0.96–0.98) | <. 001 | |
| Age | 1 (1–1) | .727 | |
| Elixhauser index | 1.02 (1.02–1.03) | <. 001 | |
| Insurance | |||
| Medicare | ref | - | |
| Medicaid | 0.98 (0.97–1) | .080 | |
| Private insurance | 0.96 (0.94–0.97) | <. 001 | |
| Self-pay | 0.89 (0.84–0.95) | <. 001 | |
| No pay | 0.79 (0.65–0.96) | .019 | |
| Other | 0.9 (0.87–0.93) | <. 001 | |
| Income quartile | |||
| Lowest (0–25) | ref | - | |
| Middle Low (26–50) | 0.99 (0.97–1) | .066 | |
| Middle High (51–75) | 0.99 (0.97–1) | .170 | |
| Highest (76–100) | 1.01 (0.99–1.03) | .228 | |
| Comorbidity | |||
| Hypertension | 0.82 (0.81–0.84) | <. 001 | |
| Hyperlipidemia | 0.82 (0.81–0.83) | <. 001 | |
| Angina | 0.93 (0.89–0.97) | .002 | |
| CAD | 1.01 (1–1.03) | .029 | |
| CHF | 0.99 (0.97–1.02) | .645 | |
| History of MI | 0.88 (0.87–0.9) | <. 001 | |
| AICD | 1.11 (1.08–1.14) | <. 001 | |
| Cardiogenic shock | 1.46 (1.45–1.48) | <. 001 | |
| Endocarditis | 0.95 (0.94–0.97) | <. 001 | |
| Chronic kidney disease | 0.99 (0.98–1) | .055 | |
| Chronic pulmonary disease | 0.94 (0.93–0.96) | <. 001 | |
| Peripheral vascular disease | 1.11 (1.09–1.14) | <. 001 | |
| Chronic liver disease | 1.08 (1.04–1.11) | <. 001 | |
| Diabetes | 1.07 (1.05–1.09) | <. 001 | |
| Anemia | 0.95 (0.94–0.96) | <. 001 | |
| Coagulopathy | 0.99 (0.98–1.01) | .329 | |
| Frailty | 1.35 (1.32–1.39) | <. 001 | |
| High BMI (30+) | 0.88 (0.86–0.89) | <. 001 | |
| History of stroke | 1.41 (1.38–1.45) | <. 001 | |
| Bed size | |||
| Small | ref | - | |
| Medium | 1.31 (0.95–1.8) | .097 | |
| Large | 1.15 (0.86–1.55) | .341 | |
| Ownership | |||
| Government | ref | - | |
| Nonprofit | 0.89 (0.78–1.01) | .079 | |
| Private | 0.71 (0.51–1) | .048 | |
| Teaching status | |||
| Non-teaching | ref | - | |
| Teaching | 1.05 (0.84–1.31) | .652 | |
OHT, orthotopic heart transplantation; VAD, ventricular assist device; IRR, incidence rate ratio; CABG, coronary artery bypass grafting.
After adjusting for patient and hospital-level factors, the VAD cohort had more than a two-fold greater odds mortality compared with patients treated with OHT (AOR = 2.22, CI 1.67–2.97, P < .001) (Table V). Increasing age was slightly associated with mortality (OR = 1.02, CI 1.01–0.03, <0.001). Sex, Elixhauser Index, income quartile, type of insurance, and hospital type were not predictors of death. Patients with a history of stroke, cardiogenic shock, peripheral vascular disease, and coagulopathy had increased odds of inpatient mortality (Table V).
Table V.
Outcomes of index admission mortality adjusted for demographics, comorbidities, and disease severity.
| Patient-level covariates | OR | P value | |
|---|---|---|---|
| OHT | ref | ||
| VAD | 2.22 (1.67–2.97) | <. 001 | |
| Previous VAD | 0.85 (0.61–1.19) | .353 | |
| Female | 0.95 (0.71–1.28) | .748 | |
| Age | 1.02 (1.01–1.03) | <. 001 | |
| Elixhauser Index | 0.96 (0.89–1.03) | .284 | |
| Insurance | |||
| Medicare | ref | ||
| Medicaid | 0.84 (0.53–1.33) | .453 | |
| Private insurance | 0.83 (0.63–1.09) | .182 | |
| Self-pay | 1.88 (0.67–5.27) | .230 | |
| Other | 0.82 (0.4–1.68) | .593 | |
| Income quartile | |||
| Lowest (0–25) | ref | ||
| Middle low (26–50) | 1.01 (0.71–1.43) | .957 | |
| Middle high (51–75) | 0.89 (0.62–1.28) | .542 | |
| Highest (76–100) | 1.07 (0.76–1.51) | .700 | |
| Comorbidity | |||
| Hypertension | 0.5 (0.33–0.78) | .002 | |
| Hyperlipidemia | 0.36 (0.25–0.51) | <. 001 | |
| Angina | 1.17 (0.4–3.4) | .778 | |
| CAD | 0.93 (0.68–1.26) | .637 | |
| CHF | 0.51 (0.31–0.83) | .007 | |
| History of MI | 0.56 (0.34–0.9) | .018 | |
| AICD | 0.14 (0.04–0.44) | .001 | |
| Cardiogenic shock | 1.48 (1.16–1.9) | .002 | |
| Endocarditis | 0.87 (0.6–1.26) | .457 | |
| Chronic kidney disease | 0.90 (0.67–1.22) | .498 | |
| Chronic pulmonary disease | 0.77 (0.49–1.22) | .266 | |
| Peripheral vascular disease | 1.99 (1.25–3.19) | .004 | |
| Chronic liver disease | 1.44 (0.59–3.5) | .426 | |
| Diabetes | 0.99 (0.56–1.72) | .959 | |
| Anemia | 0.63 (0.49–0.81) | <. 001 | |
| Coagulopathy | 2.25 (1.74–2.91) | <. 001 | |
| Frailty | 1.07 (0.59–1.93) | .820 | |
| High BMI (30+) | 1.05 (0.69–1.59) | .822 | |
| History of stroke | 5.17 (3.42–7.81) | <. 001 | |
| Bed size | |||
| Small | ref | ||
| Medium | 0.67 (0.24–1.83) | .431 | |
| Large | 0.76 (0.31–1.85) | .543 | |
| Ownership | |||
| Government | ref | ||
| Nonprofit | 0.95 (0.67–1.34) | .754 | |
| Private | 0.78 (0.16–3.89) | .761 | |
| Teaching status | |||
| Teaching | 2.10 (0.85–5.20) | .110 | |
OHT, orthotopic heart transplantation; VAD, ventricular assist device; OR, odds risk; CABG, coronary artery bypass grafting.
Costs of index hospitalization associated with VAD implantation were 22% greater compared with OHT hospitalizations ($21,929, CI $17,836–$26,392, P < .001) (Table VI). Costs were not affected by patient income quartiles or type of insurance. Comorbidities associated with increased costs included history of stroke ($28,768, CI $19,279–$39,002, P < .001), cardiogenic shock ($28,262, CI $24,251–$32,400, P < .001), and automatic implantable cardioverter defibrillator (AICD) ($16,239, CI $7,834–$25,290, P < .001).
Table VI.
Adjusted cost of care after OHT versus VAD during index admission.
| Patient-level covariates | Cost | P value | |
|---|---|---|---|
| OHT | ref | ||
| VAD | $21,929 ($17,836–$26,163) | <. 001 | |
| Previous VAD | $−2,332 ($−5,988 to $1,462) | .225 | |
| Female | $143 ($−3,216 to $3,617) | .934 | |
| Age | $16 ($−113 to $145) | .807 | |
| Elixhauser Index | $1,843 ($934–$2,760) | <. 001 | |
| Insurance | |||
| Medicare | ref | ||
| Medicaid | $−1,660 ($−6621 to $3,559) | .526 | |
| Private insurance | $−2,845 ($−5,902 to $310) | .077 | |
| Self-pay | $−5,475 ($−19,523 to $10,978) | .492 | |
| No pay | $17,782 ($−38,717 to $125,041) | .622 | |
| Other | $−8,358 ($−15,679 to $−412) | .040 | |
| Income quartile | |||
| Lowest (0–25) | ref | ||
| Middle low (26–50) | $−1,221 ($−5,219 to $2,942) | .560 | |
| Middle high (51–75) | $1,198 ($−2,975 to $5,547) | .579 | |
| Highest (76–100) | $1,986 ($−2,335 to $6,495) | .373 | |
| Comorbidity | |||
| Hypertension | $−10,941 ($−14,503 to $−7,233) | <. 001 | |
| Hyperlipidemia | $−10,704 ($−13,610 to $−7,703) | <. 001 | |
| Angina | $3,661 ($−7,395 to $16,014) | .532 | |
| CAD | $3,194 ($−543 to $7,069) | .095 | |
| CHF | $3,658 ($−2,861 to $10,607) | .278 | |
| History of MI | $−5,192 ($−9,492 to $−691) | .024 | |
| AICD | $16,239 ($7,834–$25,290) | <. 001 | |
| Cardiogenic shock | $28,262 ($24,251–$32,400) | <. 001 | |
| Endocarditis | $−3,676 ($−7,634 to $449) | .080 | |
| Chronic kidney disease | $−442 ($−4,007 to $3,254) | .812 | |
| Chronic pulmonary disease | $−3,121 ($−7,598 to $1,570) | .189 | |
| Peripheral vascular disease | $8,328 ($1,478–$15,632) | .016 | |
| Chronic liver disease | $10,160 ($−1,285 to $22,912) | .084 | |
| Diabetes | $334 ($−5,508 to $6,530) | .913 | |
| Anemia | $−779 ($−3,839 to $2,376) | .624 | |
| Coagulopathy | $4,547 ($1,280–$7,917) | .006 | |
| Frailty | $19,327 ($10,414–$28,949) | <. 001 | |
| High BMI (30+) | $−3,656 ($−7,898 to $779) | .105 | |
| History of stroke | $28,768 ($19,279–$39,002) | <. 001 | |
| Bed size | |||
| Small | Ref | ||
| Medium | $98,797 ($54,308–$155,980) | <. 001 | |
| Large | $43,874 ($14,091–$81,325) | .002 | |
| Ownership | |||
| Government | ref | ||
| Nonprofit | $−18,178 ($−26,096 to $−9,431) | <. 001 | |
| Private | $−64,956 ($−74,417 to $−52,202) | <. 001 | |
| Teaching status | |||
| Non-teaching | ref | ||
| Teaching | $11,507 ($−6,955 to $33,570) | .238 | |
OHT, orthotopic heart transplantation; VAD, ventricular assist device; CABG, coronary artery bypass grafting.
Compared with OHT, patients undergoing VAD implantation had an increased rate of postoperative supraventricular tachycardia (SVT)/atrial fibrillation (48% versus 37%, P < .001), myocardial infarction (6.9% versus 2.3%, P < .001), sepsis (12% versus 8.7%, P = .024), and urinary tract infections (17% versus 9%, P < .001) (Table VII). In contrast, patients having OHT had greater rates of postoperative pneumonia (9.3% versus 7.5%, P < .0155), and pneumothorax (4.6% versus 1.89%, P < .001).
Table VII.
Complications during index hospital stay.
| OHT (%) | VAD (%) | P value | |
|---|---|---|---|
| Valvular insufficiency | 126 (2.3) | 112 (1.7) | .235 |
| Puncture | 94 (1.7) | 112 (1.7) | .920 |
| Hemorrhage | 740 (13.6) | 1071 (16.1) | .122 |
| Hematoma | 269 (5.0) | 295 (4.2) | .540 |
| Dissection | 4.2 (0.1) | 3.4 (0.1) | .680 |
| Stroke | 70 (1.3) | 116 (1.7) | .257 |
| Supraventricular tachycardia | 2023 (37) | 3229 (48) | <. 001 |
| Atrioventricular block | 195 (3.6) | 172 (2.6) | .192 |
| Shock | 1018 (19) | 1213 (19) | .572 |
| Myocardial infarction | 127 (2.3) | 463 (6.9) | <. 001 |
| Pulmonary embolism | 76 (1.4) | 132 (2.0) | .195 |
| Mural aneurysm | 18.5 (0.34) | 53 (0.80) | .139 |
| DVT | 155 (2.9) | 203 (3.0) | .768 |
| Pneumothorax | 251 (4.6) | 126 (1.89) | <. 001 |
| Pulmonary edema | 65 (1.2) | 64 (0.95) | .430 |
| Pulmonary collapse | 1168 (21) | 1437 (22) | .970 |
| Pneumonia | 506 (9.3) | 499 (7.5) | .016 |
| Sepsis | 478 (8.7) | 768 (12) | .024 |
| Wound infection | 124 (2.3) | 102 (1.5) | .076 |
| Postoperative UTI | 488 (9.0) | 1135 (17) | <. 001 |
OHT, orthotopic heart transplantation; VAD, ventricular assist device; DVT, deep ve-nous thrombosis; UTI, urinary tract infection.
Readmission
At 30 and 60 days, 1,312 (24%) and 2,499 (46%) VAD patients and 1,922 (29%) and 4,144 (62%) OHT patients were readmitted, respectively (Table VIII). Cost of readmission and duration of stay were significantly greater among VAD patients at 30 days and 6 months (Table VIII). Odds of readmission were not different at 30 days; however, readmission at 6 months was greater among the VAD cohort (OR = 1.60, CI 1.38–1.84, P < .001) after adjustment (Table IX). Patients with a history of hypertension, myocardial infarction, AICD, cardiogenic shock, and peripheral vascular disease had an increased odds of 6-month readmission. Age, type of insurance, and income quartile did not affect odds of readmission (Table X).
Table VIII.
Readmission outcomes after VAD versus OHT.
| OHT | VAD | P value | |
|---|---|---|---|
| 30-Day readmission no. | 1312 (24) | 1922 (29) | .005 |
| Cost | $21,586 ($18557–$24,614) | $29,115 ($24,285–$33,944) | <. 001 |
| Duration of stay | 6.81 (6.01–7.61) | 10.1 (8.92–11.2) | .018 |
| 6-Month readmission no. | 2499 (46) | 4144 (62) | <. 001 |
| Cost | $20,144 ($17,705–$22,583) | $34,878 ($30,552–$39,204) | .011 |
| Duration of stay | 6.28 (5.66 –6.89) | 10.4 (9.45.–11.2) | .025 |
OHT, orthotopic heart transplantation; VAD, ventricular assist device.
Table IX.
Readmission: 30-day and 6-month outcomes adjusted for demographics, comorbidities, and disease severity.
| 30-Day readmission | 6-Month readmission | ||||
|---|---|---|---|---|---|
| Patient-level covariates | OR | P value | OR | P value | |
| OHT | ref | ref | |||
| VAD | 1.02 (0.88–1.2) | .762 | 1.6 (1.38–1.84) | <. 001 | |
| Previous VAD | 0.89 (0.74–1.07) | .208 | 0.83 (0.7–0.98) | .026 | |
| Female | 1.11 (0.94–1.3) | .228 | 1.42 (1.22–1.65) | <. 001 | |
| Age | 1 (1–1.01) | .450 | 1 (0.99–1) | .290 | |
| Elixhauser Index | 1.05 (1.01–1.09) | .020 | 1.04 (1–1.08) | .067 | |
| Insurance | |||||
| Medicare | ref | ref | |||
| Medicaid | 0.84 (0.65–1.08) | .165 | 1.01 (0.8–1.26) | .953 | |
| Private insurance | 0.88 (0.75–1.03) | .103 | 0.93 (0.8–1.07) | .290 | |
| Self-pay | 0.92 (0.42–2.01) | .833 | 0.66 (0.33–1.34) | .256 | |
| No Pay | 2.91 (0.17–48.31) | .457 | - | - | |
| Other | 0.6 (0.38–0.93) | .021 | 0.62 (0.43–0.89) | .009 | |
| Income quartile | |||||
| Lowest (0–25) | ref | ref | |||
| Middle low (26–50) | 1.08 (0.89–1.31) | .429 | 1.02 (0.85–1.22) | .839 | |
| Middle high (51–75) | 0.95 (0.78–1.16) | .624 | 0.89 (0.74–1.06) | .196 | |
| Highest (76–100) | 0.82 (0.67–1) | .048 | 0.84 (0.7–1) | .051 | |
| Comorbidity | |||||
| Hypertension | 0.8 (0.66–0.98) | .031 | 0.85 (0.71–1.01) | .070 | |
| Hyperlipidemia | 1.02 (0.87–1.2) | .793 | 1.14 (0.98–1.31) | .086 | |
| Angina | 1.41 (0.85–2.35) | .188 | 1.25 (0.76–2.07) | .381 | |
| CAD | 1.04 (0.87–1.25) | .637 | 1.12 (0.95–1.32) | .173 | |
| CHF | 0.91 (0.67–1.25) | .572 | 0.9 (0.68–1.19) | .453 | |
| History of MI | 1.05 (0.84–1.31) | .660 | 1.06 (0.86–1.3) | .579 | |
| AICD | 1.21 (0.85–1.71) | .294 | 1.02 (0.73–1.42) | .917 | |
| Cardiogenic shock | 1 (0.87–1.16) | .998 | 0.89 (0.78–1.02) | .096 | |
| Endocarditis | 0.83 (0.68–1.02) | .077 | 0.83 (0.69–0.99) | .04 | |
| Chronic kidney disease | 1.02 (0.86–1.21) | .834 | 1.03 (0.88–1.21) | .74 | |
| Chronic pulmonary disease | 1.16 (0.93–1.45) | .184 | 1.28 (1.04–1.59) | .022 | |
| Peripheral vascular disease | 0.78 (0.57–1.09) | .142 | 0.83 (0.62–1.11) | .201 | |
| Chronic liver disease | 0.94 (0.55–1.59) | .805 | 0.93 (0.58–1.51) | .783 | |
| Diabetes | 0.84 (0.62–1.13) | .245 | 1.12 (0.86–1.46) | .416 | |
| Anemia | 1.05 (0.91–1.21) | .523 | 1.17 (1.02–1.33) | .021 | |
| Coagulopathy | 0.83 (0.71–0.97) | .021 | 0.83 (0.72–0.95) | .007 | |
| Frailty | 1.42 (1–2.01) | .052 | 1.55 (1.09–2.2) | .015 | |
| High BMI (30+) | 0.94 (0.76–1.17) | .571 | 0.99 (0.81–1.21) | .934 | |
| History of stroke | 0.78 (0.53–1.16) | .226 | 0.69 (0.49–0.97) | .031 | |
| Bed size | |||||
| Small | |||||
| Medium | 2.75 (1.49–5.07) | .001 | 2.72 (1.55–4.77) | <. 001 | |
| Large | 1.95 (1.12–3.42) | .019 | 2.14 (1.3–3.51) | .003 | |
| Classification | |||||
| Government | |||||
| Nonprofit | 0.92 (0.75–1.11) | .378 | 0.9 (0.73–1.1) | .287 | |
| Private | 0.8 (0.31–2.07) | .651 | 0.68 (0.28–1.65) | .393 | |
| Teaching status | |||||
| Non-teaching | |||||
| Teaching | 1.28 (0.86–1.91) | .223 | 0.95 (0.64–1.4) | .79 | |
OHT, orthotopic heart transplantation; VAD, ventricular assist device; CABG, coronary artery bypass grafting.
Table X.
Primary causes of readmission by replacement modality.
| 30-Day readmission | 6-Month readmission | ||||||
|---|---|---|---|---|---|---|---|
| OHT (n=1,312) | VAD (n=1,922) | OHT (n=2,499) | VAD (n=4,144) | ||||
| Complications of transplanted heart | 387 (31.0%) | Acute on chronic systolic heart failure | 188 (9.7%) | Complications of transplanted heart | 586 (23.5%) | Infection because of VAD | 322 (7.8%) |
| Postoperative infection (septicemia, influenza, pneumonia, C diff colitis, UTI) | 101 (7.7%) | Gastrointestinal tract bleed | 151 (7.8%) | Postoperative infection (septicemia, influenza, pneumonia, C diff colitis, UTI) | 197 (8.0%) | Acute on chronic systolic heart failure | 306 (7.5%) |
| Acute kidney failure | 51 (3.9%) | VAD mechanical complications | 102 (5.9%) | Acute kidney failure | 112 (4.5%) | VAD mechanical complications | 361 (7.9%) |
| Pulmonary embolism and infarction | 31 (2.4%) | Paroxysmal ventricular tachycardia | 112 (5.8%) | Pulmonary embolism | 36 (1.4%) | Paroxysmal ventricular tachycardia | 174 (4.2%) |
| Seroma | 30 (2.2%) | Infection due to VAD | 70 (3.6%) | Seroma | 33 (1.3%) | Gastrointestinal tract bleed | 313 (7.6%) |
| Cardiac dysrhythmias | 16 (2.1%) | Abnormal coagulation profile | 68 (3.5%) | Other pulmonary embolism and infarction | 36 (1.4%) | Chronic ischemic heart disease | 115 (2.8%) |
| Disruption of surgical wound | 22 (1.7%) | Other postoperative infection | 58 (3%) | Complications of transplanted kidney | 29 (1.2%) | Abnormal coagulation profile | 94 (2.3%) |
| Gastrointestinal bleed | 25 (1.9%) | Unspecified transient cerebral ischemia | 32 (1.7%) | Disruption of surgical wound | 24 (0.9%) | Unspecified cerebral artery occlusion | 55 (1.3%) |
OHT, orthotopic heart transplantation; VAD, ventricular assist device; C diff, clostridium difficile; UTI, urinary tract infection.
Discussion
Although OHT remains the definitive treatment for end-stage HF, VAD technology has improved outcomes for patients with advanced HF in the past decade and has become an established treatment modality.5–8 Outcomes with such pumps have improved because of advances in technology, operative technique, and post-implantation management. With increasing experience in the use of VAD as DT, morbidity and mortality for patients is expected to approach that of OHT, thus obviating the need to utilize limited organ resources. The present study compared trends in the management of advanced HF on a national level. More than a 5-year span from 2010 to 2014, our analysis demonstrated increased utilization of VAD implantation with the comparable cost of implantation and rates of 30-day and 6-month readmission between OHT and VAD.
The incidence of VAD implantation nearly doubled from 2010 to 2014; whereas the incidence of OHT increased only marginally, a trend that is consistent with previous analyses of national administrative data.1 In our analysis, index hospitalization mortality after VAD implantation averaged 10% during the years 2010–2014, indicating that procedure-related mortality has continued to decrease after Mulloy et al.14 reported its decrease from 40% to 18% from 2005 to 2009.1,14 These decreases likely reflect the improvements being made in the field of mechanical circulatory assistance, postoperative and implantation follow-up, and the evolving demographic and comorbidity characteristics of patients.15
From 2010 to 2014, the OHT index hospitalization cost increased by 47% from $140,851 to $206,793, a trend that was consistent with Mulloy et al.,14 who reported a 40% increase from $120,413 to $168,576. Our VAD index hospitalization cost remained stable during the study period at a mean of $213,000. Our estimates of the cost of implantation were consistent with more recent studies, including the estimates of Slaughter et al.16 of $193,812 domestically. The decreased hospitalization costs of VAD implantation during the past decade are largely attributable to the improved technology of continuous flow pumps, which afford better reliability, lesser rates of pump exchange, and lesser rates of infectious complications.16
A major contributor to the costs of index hospitalization is the duration of stay. Average hospitalization of the index admission for VAD implantation was 36 days compared with 35 days in patients receiving OHT. Notably, our analysis indicated that the average stay for VAD implantation decreased from 38 days to 35 days from 2010 to 2014, whereas stay for OHT increased from 31 days to 42 days. Our 36-day average duration of stay for VAD implantation is greater than the 20-day average stay reported in previous analyses, which had smaller sample sizes.17,18 History of mechanical circulatory support did not impact the post-operative stay in patients undergoing OHT. Despite a shorter total hospital stay, patients with OHT bridged with VAD had a similar post-operative stay compared with OHT patients without bridging. These findings are similar to other studies detailing outcomes of BTT with VAD.19–21 The lesser pre-operative period in the OHT after BTT group may reflect systemic differences in communication and follow-up practices enabling patient admission just before the transplant.
Our analysis demonstrated that postoperative bleeding, cardiac complications, including arrhythmias and MI, pneumothorax, and infectious complications after VAD implantation were common. Our findings were similar to previous literature examining post-implantations complications.2,14, 16, 22–24 Akhter et al.25 found postoperative bleeding led to an additional 3 days, infections an additional 5 days, and cardiac complications an additional 7 days in hospital stay.25 Slaughter et al.16 estimated that each of the post-implantation complications led to an incremental increase in cost ranging between $22,000 and $53,000. These figures high-light the need for meticulous perioperative management and the development and dissemination of best practices to decrease costs and durations of hospital stay.
All-cause HF readmission at 30 days has been estimated to be as low as 6% in previous studies.26 Compared with patients who underwent OHT, patients receiving a VAD had a 5% and 16% greater rate of readmission after 30 days and 6 months, respectively. Our 29% cumulative incidence of readmission within 30 days after dis-charge after VAD implantation is less than the 44% reported by Akhter et al.25 and comparable to the 22% reported by Hasin et al.27 Our 62% 6-month readmission incidence after VAD implantation is also similar to the 55.6% reported by Hasin et al.27 and the 79% reported by Forest et al.28 Compared with the OHT cohort, average cost of readmission was 34% ($7,529) greater at 30 days and 73% ($14,734) greater at 6 months among VAD patients. As reported in similar studies,22,24, 25 common causes of readmission in VAD patients included mechanical complications and infections of the VAD, HF, cardiac arrhythmias, and gastrointestinal bleeds. To curtail VAD readmission rates and associated costs, it is important to evaluate timing and causes of readmission in order to implement cost-effective measures to address common causes and focus on prevention. Given the projected increases in VAD utilization and the limited pool of transplant donors, these measures are essential to the viability of such therapy in the era of value-based health care delivery.
Limitations
This study was subject to the limitations consistent with retrospective, aggregated, administrative data. The identification of diseases and procedures was limited to variables that were available in the registry based on ICD-9 coding. The NRD is limited to inpatient admissions only. Duration of stay post-operatively is not separated from total duration of stay for years before 2014 included in the analysis. VAD model or mechanism and subsequent management were not systematically captured despite recognized differences in left versus right ventricular support, technology generation, BTT, and pharmacologic support. Furthermore, the granularity of data was limited to diagnostic and procedural data. Race, laboratory values, imaging, and medical therapy were un-available in the NRD. Complications after transplant are combined in a single complication code limiting identification of types of rejection or incidence of opportunistic infection. Although patient records were linked for multiple hospitalizations within an annual period, outpatient evaluations, emergency room visits, and access to pharmacy data were not included in the database.
In conclusion, during a 5-year span from 2010 to 2014, nation-wide rates of VAD implantation doubled and index hospitalization mortality after VAD implantation averaged 10%, nearly 4 times less than the 42% mortality reported in 2005.1 Implant hospitalization costs have also appreciably decreased in the past decade, with improved technology of continuous flow pumps. Nonetheless, 30-day and 6-month rates of VAD patient readmission remain greater than their OHT counterparts. Given the projected increases in VAD utilization and the limited transplant donor pool, further emphasis on cost containment and decreased rates of readmissions after VAD implantation is essential to the viability of such therapy in the era of value-based health care delivery.
Footnotes
Supplementary materials
Supplementary material associated with this article can be found, in the online version, at doi: 10.1016/j.surg.2018.04.013.
Presented at the 13th Annual Academic Surgical Congress.
References
- 1.Mulloy DP, Bhamidipati CM, Stone ML, Ailawadi G, Kron IL, Kern JA. Orthotopic heart transplant versus left ventricular assist device: A national comparison of cost and survival. J Thorac Cardiovasc Surg. 2013;145:566–573 discussion 573–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Tuncer ON, Kemaloğlu C, Erbasan O, Gölbaşı İ, Türkay C, Bayezid Ö. Outcomes and readmissions after continuous flow left ventricular assist device: Heartmate II versus Heartware Ventricular Assist Device. Transplant Proc. 2016;48:2157–2161. [DOI] [PubMed] [Google Scholar]
- 3.McCartney SL, Patel C, Del Rio JM. Long-term outcomes and management of the heart transplant recipient. Best Pract Res Clin Anaesthesiol. 2017;31:237–248. [DOI] [PubMed] [Google Scholar]
- 4.Jamil A, Qin H, Felius J, Saracino G, Rafael AE, MacHannaford JC, et al. Comparison of clinical characteristics, complications, and outcomes in recipients having heart transplants <65 years of age versus ≥65 years of age. Am J Cardiol. 2017;120:2207–2212. [DOI] [PubMed] [Google Scholar]
- 5.Dossow von V, Costa J, D’Ovidio F, Marczin N. Worldwide trends in heart and lung transplantation: Guarding the most precious gift ever. Best Pract Res Clin Anaesthesiol. 2017;31:141–152. [DOI] [PubMed] [Google Scholar]
- 6.Weiss ES, Allen JG, Kilic A, Russell SD, Baumgartner WA, Conte JV, et al. Development of a quantitative donor risk index to predict short-term mortality in orthotopic heart transplantation. J Heart Lung Transplant. 2012;31:266–273. [DOI] [PubMed] [Google Scholar]
- 7.Khoshbin E, Dark JH. Heart transplantation. Surgery (Oxford). 2017;35:360–364. [Google Scholar]
- 8.Kim JH, Singh R, Pagani FD, Desai SS, Haglund NA, Dunlay SM, et al. Ventricular assist device therapy in older patients with heart failure: Characteristics and outcomes. J Card Fail. 2016;22:981–987. [DOI] [PubMed] [Google Scholar]
- 9.Seco M, Zhao DF, Byrom MJ, Wilson MK, Vallely MP, Fraser JF, et al. Long-term prognosis and cost-effectiveness of left ventricular assist device as bridge to transplantation: A systematic review. Int J Cardiol. 2017;235:22–32. [DOI] [PubMed] [Google Scholar]
- 10.Bansal N, Hailpern SM, Katz R, Hall YN, Tamura MK, Kreuter W, et al. Outcomes associated with left ventricular assist devices among recipients with and without end-stage renal disease. JAMA Intern Med. 2018;178:204–209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36:8–27. [DOI] [PubMed] [Google Scholar]
- 12.Bozso SJ, Nagendran J, Gill RS, Freed DH, Nagendran J. Impact of obesity on heart and lung transplantation: Does pre-transplant obesity affect outcomes? Transplan Proc 2017; 49: 344–7. [DOI] [PubMed] [Google Scholar]
- 13.Dunlay SM, Haas LR, Herrin J, Schilz SR, Stulak JM, Kushwaha SS, et al. Use of post-acute care services and readmissions after left ventricular assist device implantation in privately insured patients. J Card Fail. 2015;21:816–823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Pawale A, Schwartz Y, Itagaki S, Pinney S, Adams DH, Anyanwu AC. Selective implantation of durable left ventricular assist devices as primary therapy for refractory cardiogenic shock. J Thorac Cardiovasc Surg. 2018;155:1059–1068. [DOI] [PubMed] [Google Scholar]
- 15.Nativi JN, Brown RN, Taylor DO, Kfoury AG, Kirklin JK, Stehlik J, et al. Temporal trends in heart transplantation from high-risk donors: Are there lessons to be learned? A multi-institutional analysis. J Heart Lung Transplant. 2010;29:847–852. [DOI] [PubMed] [Google Scholar]
- 16.Slaughter MS, Bostic R, Tong K, Russo M, Rogers JG. Temporal changes in hospital costs for left ventricular assist device implantation. J Card Surg. 2011;26:535–541. [DOI] [PubMed] [Google Scholar]
- 17.Cotts WG, McGee EC, Myers SL, Naftel DC, Young JB, Kirklin JK, et al. Predictors of hospital length of stay after implantation of a left ventricular assist device: An analysis of the INTERMACS registry. J Heart Lung Transplant. 2014;33:682–688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Taghavi S, Jayarajan SN, Komaroff E, Mangi AA. Right ventricular assist device results in worse post-transplant survival. J Heart Lung Transplant. 2016;35:236–241. [DOI] [PubMed] [Google Scholar]
- 19.Yoshioka D, Li B, Takayama H, Garan RA, Topkara VK, Han J, et al. Out-come of heart transplantation after bridge-to-transplant strategy using various mechanical circulatory support devices. Interact Cardiovasc Thorac Surg. 2017;25:918–924. [DOI] [PubMed] [Google Scholar]
- 20.Osaki S, Edwards NM, Johnson MR, Velez M, Munoz A, Lozonschi L, et al. Improved survival after heart transplantation in patients with bridge to transplant in the recent era: A 17-year single-center experience. J Heart Lung Transplant. 2009;28:591–597. [DOI] [PubMed] [Google Scholar]
- 21.Urban M, Pirk J, Dorazilova Z, Netuka I. How does successful bridging with ventricular assist device affect cardiac transplantation outcome? Interact Cardiovasc Thorac Surg 2011;13:405–9. [DOI] [PubMed] [Google Scholar]
- 22.Bueno MG, Cubero JS, Fiz SS, Ugarte Basterrechea J, Hernández Pérez FJ, Goirigolzarri Artaza J, et al. Experience with a long-term pulsatile ventricular assist device as a bridge to heart transplant in adults. Rev Esp Cardiol (Engl Ed). 2017;7:727–735. [DOI] [PubMed] [Google Scholar]
- 23.Montoya JG, Giraldo LF, Efron B, Stinson EB, Gamberg P, Hunt S, et al. Infectious complications among 620 consecutive heart transplant patients at Stanford University Medical Center. Clin Infect Dis. 2001;33:629–640. [DOI] [PubMed] [Google Scholar]
- 24.Wever-Pinzon O, Naka Y, Garan AR, Takeda K, Pan S, Takayama H. National trends and outcomes in device-related thromboembolic complications and malfunction among heart transplant candidates supported with continuous-flow left ventricular assist devices in the United States. J Heart Lung Transplant. 2016;35:884–892. [DOI] [PubMed] [Google Scholar]
- 25.Akhter SA, Badami A, Murray M, Kohmoto T, Lozonschi L, Osaki S, et al. Hospital readmissions after continuous-flow left ventricular assist device implantation: Incidence, causes, and cost analysis. Ann Thorac Surg. 2015;100:884–889. [DOI] [PubMed] [Google Scholar]
- 26.Colavecchia AC, Putney DR, Johnson ML, Aparasu RR. Discharge medication complexity and 30-day heart failure readmissions. Res Social Adm Pharm. 2017;13:857–863. [DOI] [PubMed] [Google Scholar]
- 27.Hasin T, Marmor Y, Kremers W, Topilsky Y, Severson CJ, Schirger JA, et al. Readmissions after implantation of axial flow left ventricular assist device. J Am Coll Cardiol. 2013;61:153–163. [DOI] [PubMed] [Google Scholar]
- 28.Forest SJ, Bello R, Friedmann P, Casazza D, Nucci C, Shin JJ, et al. Readmissions after ventricular assist device: Etiologies, patterns, and days out of hospital. Ann Thorac Surg. 2013;95:1276–1281. [DOI] [PubMed] [Google Scholar]