On the Same Page
"Saving blood, saving money, saving lives"
October 1, 2010
You've heard it before, but it bears repeating: Quality patient care is our first and most
Now imagine a procedure performed about 15 million times per year in U.S. hospitals
that increases the likelihood of death by 70 percent and the risk of infection by about 80
percent, and is associated with other complications such as acute respiratory distress
This procedure, performed at an annual cost of $10 billion to $15 billion, is blood
transfusion. It is the single most commonly coded procedure for hospital discharges in
the United States, according to the Agency for Healthcare Research and Quality.
Interviews with several of our own faculty suggest that improving the quality of care for
our patients by developing literature-based protocols for blood transfusion and reducing
adverse events is good medicine. And it also saves money.
So we've gathered a team of physicians, nurses and laboratory staff to develop
common protocols to ensure an evidence-based approach to blood product transfusion.
In Gainesville, this team will be led by Marc Zumberg, M.D., an associate professor of
medicine, and Philip Efron, M.D., an assistant professor of surgery and anesthesiology.
In Jacksonville, the team leaders will be David Wolfson, M.D., an assistant professor of
pathology and laboratory medicine, and Agnes Aysola, M.D., an assistant professor of
pathology and laboratory medicine.
Each year, approximately 28,000 units of red blood cells are transfused at Shands at
UF, and about 11,000 units at Shands Jacksonville. Although the direct cost of each unit
of blood is about $175, the total cost is much greater.
After precisely mapping all diagnostic, therapeutic, technical, laboratory, logistic,
administrative and informational activities involved in the transfusion of blood in real-
world surgical settings, researchers constructed an activity-based cost model capturing
all the actual direct and indirect costs of acquiring, delivering, administering and
monitoring red blood cell transfusions from the hospital perspective. (Transfusion 2010;
50:753-65) This yielded an estimate of $520 to $1,180 for the total cost per unit of
blood that reflects the complexities of real-world blood utilization, depending on the
circumstances surrounding the transfusion.
Even if the average were at the low end (e.g., $600 per unit), a reduction of only 10
percent across Shands at UF and Shands Jacksonville (i.e., about 4,000 units) would
save $2.4 million per year.
Achieving such a reduction is a realistic goal. According to data from the University
Health System Consortium, which consists of 93 peer institutions, use of red blood cell
units by Shands at UF per annual inpatient discharges is 0.87. When UHC hospitals are
sorted by numbers of inpatient discharges, all hospitals of our size or larger have lower
rates of red blood cell use. For example, the rates for Barnes-Jewish (Washington
University), Vanderbilt University and University of Arizona are 0.68, 0.64 and 0.56,
In addition, data reported to UHC indicate an annual use of 13,500 units of plasma by
Shands at UF and 8,200 units by Shands Jacksonville. Plasma is the liquid part of the
blood in which the blood cells are suspended. Plasma for transfusion is usually termed
fresh-frozen plasma, or FFP. FFP is commonly used for a wide variety of indications,
but review of the literature would suggest that the evidence favors plasma transfusion in
only a very limited number of clinical situations.
Specifically, a recent report from a blue-ribbon multidisciplinary guidelines panel
(Transfusion 2010;50:1227-1239), which conducted a systematic review and meta-
analysis of randomized and observational studies, concluded that FFP was indicated in
patients with blood loss requiring massive transfusion and in patients with warfarin
therapy in association with intracranial hemorrhage. The panel did not favor plasma
transfusion for other selected groups of patients.
Other opportunities become apparent after reviewing the contemporary medical
literature and listening to our hematology faculty's expert judgments. For example, it is
still common practice to attempt to correct "abnormal" laboratory values of hemostasis
prophylactically in patients with liver disease by administering blood products. But a
recent review of the literature suggests that such a practice is not supported by the
evidence. Thanks to tremendous progress in the understanding of hemostatic function
in patients with liver disease, the long-standing dogma that patients with liver disease
have a bleeding tendency that can be corrected by transfusion no longer appears to be
supported by data from both clinical and laboratory studies (Blood 2010;116:878-885).
Rather, it appears that attempts at preoperative transfusion in patients with liver disease
does not reduce, and may in fact promote, bleeding.
A recent review in Critical Care Mlledicine classified 45 studies including 272,596
patients as 1) risks outweigh benefits, 2) neutral risk and 3) benefits outweigh risks. In
42 of 45 studies, the risks of red blood cell transfusion outweighed the benefits, the risk
was neutral in two studies, and the benefits outweighed the risks in a single subgroup of
a single study (elderly patients with an acute heart attack and a hematocrit less than 30
Still, blood transfusion can be a life-saving procedure for many patients in specific
situations involving hemorrhage from trauma or surgery, or in many cases of severe
anemia in patients who are critically ill. The data are quite sobering, however. As
suggested in an editorial by Drs. Howard Corwin and Jeffrey Carson (NEJM
2007;356:1667-8): "Red-cell transfusion should no longer be regarded as "may help, will
not hurt" but, rather, should be approached as "first, do no harm."'
Yet over the past several decades, with the exception of concern about transfusion-
related infection that has been largely eliminated thanks to effective testing for hepatitis
and HIV, the practice of transfusion has grown dramatically under the "may help, can't
In 1999, the results of an important clinical trial, Transfusion Requirements in Critical
Care, were reported (NEJM 1999;340:409-17). In this randomized, controlled study
involving critically ill adults, a liberal strategy (transfuse if the hemoglobin level drops
below 10.0 g/dL) was compared with a restrictive strategy (transfuse if hemoglobin
drops below 7.0 g/dL). Patients randomly assigned to restrictive management received
54 percent fewer red-cell units than did the liberal management group, and the
restrictive strategy was found to be at least as effective as the liberal strategy with
respect to mortality. In patients who were less acutely ill or under 55 years of age, the
restrictive strategy was actually superior, in that compared with the liberal strategy it
was associated with lower mortality.
A study of children in a Pediatric Intensive Care Unit (NEJM 2007;365: 1609-1 9)
reached similar conclusions. Using multiple organ dysfunction as an endpoint, a
restrictive transfusion strategy was at least equivalent to the liberal strategy in this
outcome, and was associated with a 44 percent reduction in the number of red-cell
These findings come at a time when a quarter of a century of research demonstrates
that transfused patients, in general, have much poorer outcomes than similar
untransfused patients, and that patients who receive more transfusions do progressively
worse in a dose-dependent fashion (Transfusion 2005 45[Supplement]: 33S-40S).
More recent studies reveal that among patients who are treated for trauma, blood
transfusion increases pneumonia, acute respiratory distress syndrome and mortality (J.
Trauma 2005, 59: 19-23). Among burn patients, blood transfusions increase mortality
and infection, controlling for indices of burn severity (Crit Care Med 2006, 34: 1602-9).
The history of blood transfusion is intertwined with the history of medicine generally, and
is one of the great examples of true translational science. The first successful
transfusion of human blood was performed by Dr. James Blundell, a British obstetrician,
who treated a woman who developed postpartum hemorrhage using her husband's
blood. Through the 1800s, transfusions were only infrequently performed (e.g., to treat
conditions such as hemophilia), as many recipients died due to what we now know were
hemolytic reactions from blood group incompatibility.
The ABO blood group system was discovered by the Austrian Karl Landsteiner, M.D., in
1901 (for which he won the Nobel Prize in Medicine or Physiology in 1930), providing
the scientific basis for improving the safety of blood transfusion. In 1939, Drs.
Landsteiner, Levine and Weiner discovered the Rh blood group system. In 1961, Rh
immune globulin was commercially introduced to prevent Rh disease in newborns of
Since then, advances have occurred in the systems of infection screening, storage,
distribution and processing of blood products. Currently accepted blood transfusion
practices evolved prior to the concepts of randomized trials and clinical outcomes
studies, however, and developed all the sanctity expected of time-honored therapies.
In summary, interviews with faculty involved in blood transfusion suggest a consensus:
As is often the case in quality improvement initiatives, improving the quality of care for
our patients by developing literature-based protocols for blood transfusion and reducing
adverse events will also result in cost savings. I would add, in the spirit of the goals of
our Clinical and Translational Science Award, that research on blood transfusion has
progressed through the first two translational stages: from "T1" (discovery of blood
groups, fractionation and storage methods, etc., translated to individual patient
treatment), to "T2" (defining risk groups and hemoglobin targets for translation to
optimal clinical practice).
It is now time to progress to the "T3" phase, in which accumulated knowledge on best
practices can be disseminated to our physicians, nurses and students, so that known
best practices can be achieved across our health-care system in a manner that saves
blood, saves money and saves lives.
David S. Guzick, M.D., Ph.D.