Extracorporeal membrane oxygenation (ECMO) can be used for temporary life support for patients suffering from severe, but potentially reversible respiratory or cardiac failure when conventional treatments have failed or are likely to fail. On the basis of randomized trial data and the excellent survival that is witnessed in the neonatal group, ECMO has become the standard of care.1–4
The survival rates for adults with severe acute respiratory failure treated with ECMO have improved dramatically over the past 30 years.5,6 In 2009, the investigators of the CESAR trial, a multicenter randomized-controlled study in the United Kingdom, concluded that ECMO improves survival in Acute respiratory distress syndrome (ARDS) and is cost-effective.7 In addition, the therapy proved useful in the recent H1N1 epidemic.8 Multiple studies have also suggested the potential benefit of ECMO in treating adults with cardiac failure.9–13 Recently, the use of ECMO as a bridging strategy for lung transplantation has shown acceptable outcomes.14 The reemergence of interest in ECMO therapy has direct implications for patient care, surgical expertise, cardiac surgical intensive care unit resource usage, and future clinical training.
The purpose of this investigation was to analyze the usage of ECMO in adults in order to identify recent trends within the United States.
The Nationwide Inpatient Sample (NIS), part of the Healthcare Cost and Utilization Project (HCUP), sponsored by the Agency for Healthcare Research and Quality (AHRQ) from 1999 to 2011 was used.15 Although all years were analyzed, the focus was on the most recent 6 years available, 2006–2011. The NIS is the largest publicly available all-payer inpatient care database in the United States. Each year contains data for approximately 8 million inpatient hospital stays from about 1,000 hospitals sampled to approximate a 20% stratified sample of all US nonfederal hospitals. All discharges from sampled hospitals are included in the dataset. The NIS discharge weights were used to create national estimates for all US hospitalized patients in a given year. The NIS was used under the HCUP Data Use Agreement. All data in the NIS was previously de-identified before our use. This study met exemption criteria from Institutional Review Board approval at Yale University.
Calculating Usage Rates, Survival Rates, and Costs
The NIS uses procedure codes from the International Classification of Diseases, 9th edition, Clinical Modification, Volume 3. Patients who received ECMO were identified using the code 39.65. All patients with ≥18 years of age were included. Usage rates were calculated as cases of adult patients who received ECMO per million adult inpatient stays.
The survival rate was defined as the percentage of adult patients who received ECMO during hospitalization and survived to discharge. Hospitals were categorized by bed size and volume of adult ECMO cases. The NIS categorizes hospitals into three sizes (small, medium, and large) based on bed size with the definitions dependent on the teaching status and geographic region of the hospital.16 Hospitals were also divided by us into two categories depending on their volume of adult ECMO cases with high-volume hospitals being defined as performing at least six cases annually, which was based on a recommendation for ECMO centers.17
The costs were estimated using the total hospital charges for each patient adjusted using the HCUP Cost-to-Charge ratios for each hospital,18 which were obtained from the Centers for Medicare and Medicaid Services. These costs were then adjusted for inflation using the Consumer Price Index (CPI) Inpatient Hospital Services inflation multiplier.19 The total cost for each patient was divided by the length of stay to estimate the cost per day. The medians of the costs per day and the total costs per patient for each year were used to analyze trends.
Corroborative Data from ELSO Registry
Summary datasets from the ECMO Registry of Extracorporeal Life Support Organization (ELSO, Ann Arbor, MI)20 were used to corroborate trends found using the NIS. This registry collects self-reported data from over 100 centers worldwide on the use, complications, and outcomes of patients who received ECMO. Only data for adults from centers in the United States were used for comparison.
Simple linear regression analyses using summary datasets for each year were used to compare annual ECMO usage rates, survival rates, costs per day, and total costs per patient from 2006 to 2011. All data presented has been weighted using the NIS discharge weights to produce national estimates. Patient characteristics were compared using a t-test for mean age comparison and χ2 tests for comparisons of categorical variables. Two-tailed p values <5% were considered a priori to be statistically significant. Quantitative analyses were performed using Stata 12 (StataCorp LP, College Station, TX).
A comparison between adult patients who received ECMO in 2006 and 2011 is shown in Table 1. In 2006, there were 375 (95% confidence interval [CI], 200–549) cases compared with 2004 (95% CI, 889–3119) cases in 2011. There were no significant differences in the patient age, gender, payer mix, or the proportion of patients living in low-income ZIP codes. There was no significant difference in the major diagnostic category of the patients, but there were significantly more patients in 2011 compared with 2006 who were categorized as having an extreme risk of mortality (78% vs. 44%, p < 0.001).
The annual rates of ECMO cases per million adult discharges from 1999 to 2011 are shown in Figure 1. There was no significant difference from 1999 to 2007 (p for trend = 0.14), but the rate increased 433% from 11.4 (95% CI, 6.1–16.8) in 2006 to 60.9 (95% CI, 28.1–93.7) in 2011. There was a significant increase in the rate of ECMO cases per million adult discharges from 2006 to 2011, as seen in Table 2 (p for trend = 0.001).
The survival rates for 2006–2011 are shown in Figure 2. Although there is a trend towards an improvement in the survival rate from 2006 to 2011, as seen in Table 2, it is not statistically significant (p for trend = 0.14). Of the patients who survived to discharge, 16% were transferred to other acute care hospitals (Table 3). The outcomes of these patients after transfer are unknown. The survival rates of patients that were received as transfers from other acute care hospitals and then underwent ECMO, 37% (95% CI, 29–45%), do not significantly differ (p = 0.54) from the survival rates of patients that were not transfers, 39% (95% CI, 34–44%).
Most ECMO cases (90%) took place in large hospitals, as seen in Figure 3. The survival rate for medium-sized hospitals is significantly better than for large (p = 0.01) and small hospitals (p = 0.006). However, survivors from medium-sized hospitals were much more likely to be transferred to other acute care hospitals compared with survivors from large hospitals (49% vs. 11%). In addition, the majority of ECMO cases (73%) took place in high-volume centers, as seen in Figure 4. Survivors from low-volume centers were much more likely to be transferred to other acute care hospitals compared with survivors from high-volume centers (44% vs. 5%). The survival rates between the two types of centers do not significantly differ (p = 0.53).
The costs per day of hospitalization for adults who received ECMO, as well as the hospitalization costs per patient were compared from 2006 to 2011 in Figure 5. As seen in Table 2, there were no significant differences from 2006 to 2011 in the median costs per day (p for trend = 0.07) or the median costs per patient (p for trend = 0.87).
Validation of NIS Estimates
The ELSO Registry recorded a gradual increase in adult ECMO cases from 82 in 1999 to 155 cases in 2007, followed by a rapid increase to 591 in 2011. The survival rates recorded by the ELSO Registry remained mostly constant from 2006 to 2011, averaging 41%. A paired t-test shows that there is no significant difference in survival rates between the ELSO Registry and the NIS from 2006 to 2011 (p = 0.13).
There has been a sharp increase in ECMO usage in adults in the United States since 2007. Before this, the usage has been consistently low. There are likely many contributing factors to this increase, but two possible explanations are important to mention.
First, the H1N1 pandemic hit the United States in the spring of 2009, resulting in many cases of ARDS with a much higher percentage of young, healthy adults affected compared with seasonal influenza.21 During this pandemic, ECMO was effective in treating patients who developed ARDS.8 In addition, two large studies, both published in September 2009, showed favorable survival rates for adults with severe respiratory failure treated with ECMO.5,7
Second, recent studies have shown success when using ECMO to treat cardiac failure. One study showed survival benefits for patients with in-hospital cardiac arrest that received extracorporeal cardiopulmonary resuscitation versus conventional cardiopulmonary resuscitation,11 and another study showed that ECMO support rescued 40% of patients who would have otherwise died from refractory cardiogenic shock.10 In addition to the increase in adult ECMO cases for respiratory failure, there was also an increase in the use of ECMO for adults with cardiac failure around the same time.
There was a trend toward improved survival rates from 2006 to 2011, but this did not reach statistical significance. Nonetheless, there are signs of improvement. Patients in 2011 were almost twice as likely as in 2006 to be categorized as having a higher risk of mortality upon admission. This suggests that either ECMO was not as necessary for lower acuity patients or that physicians felt more comfortable using ECMO in severely ill patients in 2011 compared with 2006. In addition, many more cases in 2011 were being performed in high-volume centers compared with in 2006, a favorable trend.
An important distinction to make in the survival rates is the percentage of discharges that were transferred to other acute care hospitals. These patients are classified as survivors as the definition is survival to hospital discharge. Unfortunately, the outcome for these transferred patients is unknown. The NIS data show that the survival rates for patients who were received as transfers from other acute care hospitals and then underwent ECMO is approximately equal to the survival rates for patients that were not received as transfers. Therefore, high transfer rates inflate the survival rates. Medium-sized hospitals are shown here to have a higher survival rate than large hospitals, but when the higher transfer rates are factored in (calculations not shown), the two types of hospitals have very similar success rates. The high transfer rates at small hospitals and low-volume hospitals show that their success rates are even lower than their already low survival rates. This provides further evidence supporting the ELSO recommendation that ECMO centers perform at least six cases annually.17
The survival rates shown here are lower than seen in most other studies, as well as in the ELSO Registry (although this difference was not statistically significant). This is likely because of the NIS representing all patients in the United States who received ECMO. Research studies tend to take place at big centers with well-established ECMO protocols, and the ELSO Registry receives self-reported data from a consistent group of large ECMO centers. When looking at individual cases in the NIS, there are many examples of ECMO being used as a heroic final effort in patients that were not likely to survive with or without ECMO. Appropriate patient selection is a key to good outcomes and comes from experience.
There have been consistent improvements in survival rates dating back to its use in the 1970s. For adults treated with ECMO for severe acute respiratory failure, the first randomized clinical trial in 1979 had a survival rate of only 9.5%,2 whereas the CESAR trial in 2009 had a survival rate of 63%.7 Similarly, there has been a tremendous improvement in survival rates for patients receiving ECMO for cardiac failure. From 1990 to 2007, the survival rates for adults who suffered cardiac arrest and were treated with ECMO increased from 30% to 59%.12 Regardless of the etiology, the survival rates will likely continue to improve due to advancements in technology, a decrease in the frequency of complications such as clotting or oxygenator failure, an improved ability to support critically ill patients, and most importantly, an improved ability to treat the underlying conditions affecting these patients.
The hospitalization costs for patients who received ECMO did not change from 2006 to 2011. The trends are more important than the values presented, since for each patient, the number represents the costs of all services and procedures received during hospitalization, not all of which pertained to ECMO. This analysis should not be used to draw any conclusions about the cost effectiveness of ECMO because they incorporate costs for other services and do not include data on length of survival or quality of life after treatment. The best cost effectiveness study to date was conducted in the UK as part of the CESAR trial, which found ECMO to be cost effective with a predicted cost per quality-adjusted-life-year (QALY) of £19,252 (~$31,000).7
Although the NIS has been used many times in estimating national trends and is well validated, it is possible that the sample over- or underrepresents patients receiving ECMO. As there are few ECMO cases relative to the total number of inpatient stays, a large center could bias estimates. Also, the NIS compiles its data from billing information which can be misleading and even sometimes incorrect. Nonetheless, we are confident in our conclusions because the NIS survey design was accounted for in our statistical analyses and the trends found match those seen in the ELSO Registry.
We have witnessed a significant increase in ECMO usage in adult patients in recent years with a trend towards improved survival rates and no increases in costs. The increase in usage is likely to continue with improvements in survival rates as the technology advances and our ability to support critically ill patients and cure their underlying conditions improves.
We would like to thank our staff statistician, Elsa Cuprak, for ensuring the accuracy of our statistical analysis. We would like to acknowledge all of the HCUP Data Partners that contribute to HCUP (http://www.hcup-us.ahrq.gov/db/hcupdatapartners.jsp). We would also like to thank the Extracorporeal Life Support Organization (ELSO, Ann Arbor, MI) for providing us with data from the ELSO Registry.
1. Hill JD, O’Brien TG, Murray JJ, et al. Prolonged extracorporeal oxygenation for acute post-traumatic respiratory failure (shock-lung syndrome). Use of the Bramson membrane lung. N Engl J Med. 1972;286:629–634
2. Zapol WM, Snider MT, Hill JD, et al. Extracorporeal membrane oxygenation
in severe acute respiratory failure. A randomized prospective study. JAMA. 1979;242:2193–2196
3. Bartlett RH. Extracorporeal life support: history and new directions. Semin Perinatol. 2005;29:2–7
4. Bartlett RH, Roloff DW, Cornell RG, Andrews AF, Dillon PW, Zwischenberger JB. Extracorporeal circulation in neonatal respiratory failure: a prospective randomized study. Pediatrics. 1985;76:479–487
5. Brogan TV, Thiagarajan RR, Rycus PT, Bartlett RH, Bratton SL. Extracorporeal membrane oxygenation
in adults with severe respiratory failure: a multi-center database. Intensive Care Med. 2009;35:2105–2114
6. Lewandowski K, Rossaint R, Pappert D, et al. High survival rate in 122 ARDS patients managed according to a clinical algorithm including extracorporeal membrane oxygenation
. Intensive Care Med. 1997;23:819–835
7. Peek GJ, Mugford M, Tiruvoipati R, et al.CESAR trial collaboration. Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation
for severe adult respiratory failure (CESAR): a multicentre randomised controlled trial. Lancet. 2009;374:1351–1363
8. Noah MA, Peek GJ, Finney SJ, et al. Referral to an extracorporeal membrane oxygenation
center and mortality among patients with severe 2009 influenza A(H1N1). JAMA. 2011;306:1659–1668
9. Massetti M, Neri E, Buklas D, et al. Repair of aortic leaflet prolapse: the “sliding leaflet technique”. Ann Thorac Surg. 2005;79:1787–1789
10. Combes A, Leprince P, Luyt CE, et al. Outcomes and long-term quality-of-life of patients supported by extracorporeal membrane oxygenation
for refractory cardiogenic shock. Crit Care Med. 2008;36:1404–1411
11. Chen YS, Lin JW, Yu HY, et al. Cardiopulmonary resuscitation with assisted extracorporeal life-support versus conventional cardiopulmonary resuscitation in adults with in-hospital cardiac arrest: an observational study and propensity analysis. Lancet. 2008;372:554–561
12. Cardarelli MG, Young AJ, Griffith B. Use of extracorporeal membrane oxygenation
for adults in cardiac arrest (E-CPR): a meta-analysis of observational studies. ASAIO J. 2009;55:581–586
13. Mendiratta P, Wei JY, Gomez A, et al. Cardiopulmonary resuscitation requiring extracorporeal membrane oxygenation
in the elderly: a review of the Extracorporeal Life Support Organization registry. ASAIO J. 2013;59:211–215
14. Hoopes CW, Kukreja J, Golden J, Davenport DL, Diaz-Guzman E, Zwischenberger JB. Extracorporeal membrane oxygenation
as a bridge to pulmonary transplantation. J Thorac Cardiovasc Surg. 2013;145:862–867
15. HCUP Nationwide Inpatient Sample (NIS). Healthcare Cost and Utilization Project (HCUP). 1999–2011. Rockville, MD Agency for Healthcare Research and Quality
16. NIS Database Documentation. HCUP Nationwide Inpatient Sample (NIS). Healthcare Cost and Utilization Project (HCUP). Rockville, MD Agency for Healthcare Research and Quality http://www.hcup-us.ahrq.gov/db/nation/nis/nisdbdocumentation.jsp
17. ELSO Guidelines for ECMO Centers. . Extracorporeal Life Support Organization (ELSO) 2010 Ann Arbor, MI
18. HCUP Cost-to-Charge Ratio Files (CCR). Healthcare Cost and Utilization Project (HCUP). 2006–2011 Rockville, MD Agency for Healthcare Research and Quality
19. Consumer Price Indexes (CPI) for 2006–2011. United States Bureau of Labor Statistics. http://www.bls.gov/cpi/#data
. Accessed March 4, 2013
20. ECLS Registry Report, United States Summary. 2013. Extracorporeal Life Support Organization (ELSO). Ann Arbor, MI http://www.elso.org
21. Jain S, Kamimoto L, Bramley AM, et al.2009 Pandemic Influenza A (H1N1) Virus Hospitalizations Investigation Team. Hospitalized patients with 2009 H1N1 influenza in the United States, April-June 2009. N Engl J Med. 2009;361:1935–1944