A study examining the relationship between blood requirements and outcomes in patients requiring extracorporeal membrane oxygenator (ECMO) support has shown that the total number of blood units transfused is predictive of mortality. Researchers also found that pediatric ECMO patients had a substantially lower mortality than adults.
“The bottom line is that 37% of the pediatric ECMO patients and 70% of the adult ECMO patients didn’t leave the hospital alive,” said Steven M. Frank, MD, professor of anesthesiology and critical care and director of the blood management program at the Johns Hopkins University School of Medicine, in Baltimore. “Given that the amount of transfused blood was an independent predictor of mortality, effective blood management strategies may be beneficial in ECMO patients.”
As Dr. Frank reported, patients requiring ECMO support are often the sickest and consume a large amount of health care resources. Reported ECMO mortality ranges from 26% to 59% in neonates, 43% to 59% in pediatric patients and 43% to 71% in adult patients, depending on the type of ECMO support used and underlying comorbidities (Am J Respir Crit Care Med 2015;191:894-901).
“When we looked at our blood management data, we found that a huge amount of blood goes to ECMO patients, which is the ultimate form of life support,” Dr. Frank said. “We saw that pediatrics especially were using huge amounts of blood—red cell, plasma and platelets—so, we wanted to see how these patients end up doing.”
To identify risk-adjusted predictors of clinical outcomes with ECMO, Dr. Frank and his colleagues retrospectively analyzed data for pediatric (n=102) and adult (n=73) ECMO patients at a tertiary care academic medical center between 2010 and 2015. The researchers analyzed age, sex, total number of blood units transfused, cardiac failure and respiratory disease to determine predictors of morbidity and mortality by univariate and multivariate analyses. Morbidity analysis consisted of 12 different events, including infection, thrombosis, and renal, respiratory and ischemic events.
As Dr. Frank reported at the 2017 annual meeting of the International Anesthesia Research Society (abstract 1493), mortality was found to be higher in adult (51/73; 69.9%) versus pediatric (38/102; 37.2%) patients, although these differences may be explained by the conditions being treated (Figure).
“We found that 50% of the adults went on ECMO for cardiac failure compared with only 18% of the pediatrics,” Dr. Frank said. “I think if you’re going on ECMO for respiratory causes, you’re probably going to have a better outcome, which could explain why the pediatric patients do better. Also, pediatric patients tend to be more resilient.”
On both univariate and multivariate analyses, no clinical characteristics were significant predictors of morbidity (Table 1). Mortality, however, was increased with adult patient status (odds ratio [OR], 5.16; 95% CI, 2.41-11.55; P<0.0001) and total blood units (OR, 1.010; 95% CI, 1.004-1.016; P=0.002) on multivariate data According to the study’s authors, the pediatric ECMO mortality rate was lower than that reported in a recent review, whereas the adult ECMO mortality rate was within the reported range.
Table 1. Predictors of Morbidity in ECMO Patients | |||
Univariate | Multivariate | ||
P Value | Odds Ratio (95% CI) | P Value | |
Adult vs. pediatric patient | 0.570 | 1.235 (0.580-2.682) | 0.587 |
Sex | 0.216 | 1.542 (0.795-3.006) | 0.200 |
Total units of blood | 0.648 | 1.002 (0.997-1.008) | 0.529 |
Cardiac failure | 0.345 | 1.234 (0.564-2.778) | 0.603 |
Respiratory disease | 0.751 | 0.768 (0.242-2.710) | 0.663 |
Table 2. Predictors of Mortality in ECMO Patients | |||
Univariate | Multivariate | ||
P Value | Odds Ratio (95% CI) | P Value | |
Adult vs. pediatric patient | <0.0001 | 5.156 (2.411-11.552) | <0.0001 |
Sex | 0.073 | 0.623 (0.318-1.210) | 0.164 |
Total units of blood | 0.012 | 1.010 (1.004-1.016) | 0.002 |
Cardiac failure | 0.034 | 1.029 (0.467-2.231) | 0.943 |
Respiratory disease | 0.196 | 1.037 (0.300-3.926) | 0.955 |
“Everybody knows that the patients who get more blood are the patients that have more complex procedures and higher severity of illness, so anytime that you look at blood retrospectively it’s associated with bad outcomes,” Dr. Frank said.
Sure enough, for every unit of red blood cells transfused, investigators observed a 1% to 2% increase in mortality.
“Whether that’s a marker of illness or whether there’s a causative effect, we can’t say,” Dr. Frank continued, “but the average adult received a median dose of 16 units of red cells. The median pediatric dose of red cells, on the other hand, was 23 units, but the length of stay was a month for the pediatrics and only 10 days for the adults.”
Findings Raise Ethical Questions
Ultimately, Dr. Frank said, these data raise an important ethical question for providers: Is it sustainable to place patients on ECMO as a last resort if 70% of them are going to die?
“If we could predict which individuals are going to die and which are going to survive, then maybe we can do a better job selecting those who should receive ECMO and all this blood,” he concluded.
Jean-Francois Pittet, MD, professor of anesthesiology and perioperative medicine at the University of Alabama at Birmingham School of Medicine, underscored the issue of overutilization of health care in the United States.
“Care is unrationed here,” Dr. Pittet said. “We spend billions of dollars in this country for surgical patients that have very poor outcomes, but we’re not doing the analysis that needs to be done. I know it’s hard to make a decision for an individual patient, but we know that 70% of patients on ECMO will die, and with cardiac failure, up to 90% will die with veno-arterial ECMO.
“Texas is the only state in the U.S. where the physician has the ability to deny treatment when it’s futile,” Dr. Pittet continued. “Elsewhere, if the family insists, the physician must do everything in his or her power. I know it’s a difficult question, but this research is an important contribution to the discussion.”
Leave a Reply
You must be logged in to post a comment.