Taking a harder line on blood transfusions

May 2007
Feature Story

Karen Lusky

Hospitals trying to zero in on the key factors that put patients at risk for blood transfusions might start by looking within.

“One of the biggest risks in the U.S. of being transfused is which doors you happen to walk through on the day of surgery,” anesthesiologist Timothy Hannon, MD, MBA, said in a recent G-2 Reports audioconference on blood management. Even within a group of surgeons or anesthesiologists, he said, you see considerable variation in blood use, with some ordering quite a bit and others very little.

In fact, he told CAP TODAY, a hospital is a “quantifiable risk factor for transfusion” for all patients, whether or not they have surgery. That’s because the hospital tends to have a “culture” for how it approaches transfusion therapy, says Dr. Hannon, medical director of the blood management program at St. Vincent Hospital, Indianapolis, and president and CEO of Strategic Healthcare Group, which offers, among other services, blood management consultation.

It has been known for some time that a restrictive transfusion strategy may be better for adult patients than a liberal strategy. Now, a new study has found that a restrictive strategy (hemoglobin threshold of 7 g/dL) for red-cell transfusion can decrease transfusion requirements without increasing adverse outcomes in stable, critically ill children (Lacroix J, et al. N Engl J Med. 2007;356:1609–1619). The mounting more-may-be-less data is why some hospitals are implementing conservative, evidence-based blood management programs.

That’s what Albert Einstein Medical Center in Philadelphia did six years ago. Since then, blood use there has dropped by more than half, with about 70 percent of the reduction occurring in surgery and 30 percent in medical practice, says Ierachmiel Daskal, MD, PhD, chairman of the Department of Pathology and Laboratory Medicine.

Overall, blood costs at the hospital are up a bit owing to the higher cost of packed red blood cells and the volume of transfusions. “But if we had not implemented the program, you could add another 30 percent to 35 percent to the budget of $3 million to $3.5 million for blood products,” Dr. Daskal says.

“There’s this notion,” says David Reardon, MD, medical director of laboratories at Lee Memorial Health System, Ft. Myers, Fla., “that it’s just blood—how problematic can it be?”

The answer is plenty problematic in certain cases, say experts, which is why casual ordering of blood is risky and expensive.

Pathologist David Jadwin, DO, of Kern Medical Center, Bakersfield, Calif., points to “emerging evidence that class for class, patients who do not receive allogeneic blood transfusion have fewer complications” and a 0.9 day shorter length of stay.

As the authors of an editorial on the new study of pediatric patients write, “Red-cell transfusion should no longer be regarded as ‘may help, will not hurt’ but, rather, should be approached as ‘first, do no harm’” (Corwin HL, Carson JL. N Engl J Med. 2007;356:1667–1669).

Change has to begin in any hospital like it did at St. Vincent Hospital: by getting buy-in at all levels. A 750-bed community hospital, St. Vincent rolled out its blood management program in 2001 by forming multidisciplinary teams to work with the biggest blood users: cardiac and orthopedic surgery, critical care, and oncology.

On the teams were physicians, nurses, administrators, laboratorians, blood bankers, pharmacists, and statisticians—“all of the people who can bring to bear solutions for more appropriate blood management,” said Dr. Hannon, who shared St. Vincent’s experience with its program during the recent audioconference.

Each team reviewed the literature in its assigned area, looking for “best-practice opportunities to reduce transfusion exposure.”

For example, says Dr. Hannon, “in orthopedic surgery we reviewed the use of predonation as a blood management strategy and came to the conclusion in 2001 that it was no longer a reasonable method to reduce transfusion exposure.” Studies showed the practice makes patients anemic before surgery, and hospitals tend not to collect enough to avoid allogeneic transfusion for high-risk surgery, he says.

Also, patients who predonate “are more likely to receive a transfusion of any type—autologous or allogeneic,” Dr. Hannon adds. “And one of the biggest safety errors leading to negative outcomes with transfusion is a clerical error.”

Today, the approach St. Vincent uses for 90 percent of its orthopedic surgical patients is “to do nothing,” Dr. Hannon says. “We prepare the person for surgery and transfuse using conservative evidence-based guidelines. About 10 percent of patients get erythropoietin [before surgery] to improve their red blood cell count.” But there’s now controversy about the risks and benefits of erythropoietin, Dr. Hannon says. Studies show, for example, that it’s associated with deep vein thrombosis. So, as with anything, you have to weigh the risks and benefits of using it, he says. (A March 2007 FDA public advisory on erythropoiesis-stimulating agents is at www.fda.gov/cder/drug/advisory/RHE2007.htm.)

Today, St. Vincent continues to monitor its blood use to maintain its “leadership position,” as Dr. Hannon puts it. When the teams began their work in 2001, they simply looked internally for improvement opportunities, he says, “not knowing if we were good, bad, or in between.” But a few years ago, St. Vincent began using blood-use benchmarks for Medicare diagnosis-related groups to keep an eye on blood use trends.

“Organizations starting out with a blood management program can use the DRG benchmarks to give them targets for improvement” and then use them for process control, Dr. Hannon suggests.

As a prelude to implementing its own blood management program, Lee Memorial Health System had Dr. Hannon and his team perform in August 2006 a weeklong audit of its blood use, to see, in part, how it stacked up against national DRG benchmarks.

In the three-campus audit, 65 percent of cases were identified as opportunities to reduce or avoid blood transfusions, according to reviewers, says Marilyn Kole, MD, system ICU director for Lee Memorial Health System, who co-presented with Dr. Hannon in the audioconference on blood management.

Based on audit results showing the hospital had room for improvement in the area of red-cell transfusions, Lee Memorial’s Dr. Reardon began educating the medical staff about the Transfusion Requirements in Critical Care, or TRICC, trial (Hebert PC, et al. N Engl J Med. 1999; 340:409–417). That study showed using a hemoglobin of 7 g/dL as a transfusion trigger in critically ill patients produced just as good or slightly better outcomes than a more liberal trigger of 10 g/dL, Dr. Reardon says.

But Lee Memorial didn’t simply reset its red-cell transfusion trigger. In fact, the key, says Dr. Kole, is that you don’t “treat the lab numbers” but rather look at the “entire patient.”

For example, the hospital’s sepsis protocol dictates giving patients with the condition five to seven liters of intravenous fluid, which artificially lowers their hemoglobin and hematocrit values. “You can try to get the hemoglobin before you hydrate the patient,” Dr. Kole says, “but then it will be high if they are dehydrated.”

“It’s a complex clinical assessment, but at least we are getting people thinking about it.”

“If the patient is improving, the last thing I want to do is add blood to the picture,” Dr. Kole continues. “If the patient develops a fever, is it from the blood or the infection?”

Recognizing the pivotal role nursing plays in managing the use of blood, Dr. Kole worked with nursing leaders at Lee Memorial. Orthopedic surgeons have told her that if they refuse to order blood when a nurse calls to say a patient’s hemoglobin is too low, they know the nursing notes will read, “Doctor refused blood transfusion.” Thus, she says, “We have to teach nurses to reset their own personal hemoglobin levels as triggers for transfusion.”

Also, “in the ICU, you have to have nurses who will feel comfortable asking the physician, ‘Why are you ordering blood for this patient?’” Dr. Kole adds.

Lee Memorial’s ultimate goal, Dr. Reardon says, is to have “close to a prospective order set” where physicians “will have to verify the criteria that justify the transfusion at the time they order it”—or document the reason for exceptions.

Decision-support computerized physician order systems for blood do reduce inappropriate usage, according to a study conducted at Brigham and Women’s Hospital, Boston. At the study’s outset, an audit of the hospital found that 73 percent of the medical staff’s routine orders for blood were inappropriate based on criteria that are “more liberal than we use at St. Vincent,” Dr. Hannon said in his audioconference. The study excluded emergency department, operating room, and pediatric transfusions (Rothschild, et al. Transfusion. 2007;47:228–239).

Simply educating the physicians about appropriate blood-ordering practices reduced the rate of inappropriate transfusion orders to 61 percent. The physician order-entry system further reduced that rate to 58 to 59 percent, says Jeffrey M. Rothschild, MD, MPH, lead investigator for the study and assistant professor of medicine at Brigham and Women’s Hospital and Harvard Medical School. The primary effect of the ordering system is to help physicians hold on to the gains they make through education, he says.

The researchers designed the computerized decision-support system, which is still in place at Brigham and Women’s, to be “fairly consistent with most national guidelines,” Dr. Rothschild says. In the study, physicians were asked when ordering a blood product to choose an indication for the transfusion. The system then checked the patient’s most recent laboratory values—hematocrit if the physician ordered red blood cells, platelet count for platelet orders, and INR for fresh frozen plasma.

The physician had to enter the diagnosis for the patient receiving blood because the “system can’t be smart enough to search the medical record for the diagnosis,” Dr. Rothschild says. Many times, a physician will give blood for a condition that developed that day, such as an acute gastrointestinal bleed for a patient admitted for a myocardial infarction, he adds.

The system’s algorithm also alerted physicians if they ordered more units of blood product than the guidelines suggest.

But the decision-support system didn’t put a stop to an order. That’s because “we can’t really provide decision-support counsel for all clinical situations,” Dr. Rothschild says. But they do identify and talk to the outliers about their transfusions.

Retrospective review, too, can help reduce the number of inappropriate transfusions. But over Dr. Jadwin’s career as a blood bank director at different hospitals, he has found clinicians typically respond to pathologists’ feedback about questionable transfusions with, “You’re just pathologists” or “You don’t have all the clinical facts.”

So Dr. Jadwin and the transfusion team at Kern Medical Center began collecting all of those clinical facts several years ago, creating spreadsheets that correlated a patient’s transfusions to his or her surgery, blood loss, clinical events, vital signs, and laboratory values. They let the data do the talking, which did the trick.

“Physicians accepted reviewer analytic comments,” Dr. Jadwin says. As a result, the teaching hospital/trauma center’s use of packed red blood cell units dropped 30 percent and use of fresh frozen plasma units fell by 45 percent compared with local non-teaching, non-trauma hospitals that, Dr. Jadwin says, “experienced significant annual increases in blood transfusions.”

He provides two common scenarios that the spreadsheet approach easily reveals:

In one example, a patient’s hemoglobin is 5.8 g/dL and the physician orders a transfusion of four units of blood. Yet after receiving the second unit, the patient’s blood pressure and heart rate become normal. “But the physician and nurse don’t use a unit-by-unit approach in transfusion,” so the patient gets two more units than clinically necessary, Dr. Jadwin says.

In another example, a patient’s hemoglobin is 5.8 g/dL, so the physician orders two units of blood, which would be expected to increase the hemoglobin by about two grams. But the post-transfusion hemoglobin is 13.8 g/dL, indicating the clinician never suspected an obvious pre-transfusion erroneous hemoglobin value.

Kern Medical Center is not using the data-entry process now, which Dr. Jadwin describes as labor-intensive. But he has since parlayed the approach into an electronic external blood utilization review service for hospitals available through Columbia Healthcare Analytics.

The optimal system is one that systematically stops inappropriate blood transfusions before they are administered. That’s the approach that Albert Einstein Medical Center uses.

Dr. Daskal, who is head of pathology there, takes a hands-on approach to keeping blood use in line with consensus-driven practices. For example, he routinely reviews the surgical list with the chief of anesthesia or the chief’s staff to identify blood requests and potential problem cases.

Then he himself contacts the surgeons whose requests deviate from the hospital’s minimal surgical blood ordering system. Most of the inappropriate requests, he says, reflect a “cultural phenomenon where a surgeon wants a safety net.”

And, of course, if a patient is actively bleeding in surgery, the blood bank responds immediately and sends more blood to the OR.

“Sometimes we go in the other direction… If someone has a 7.2 hemoglobin going into surgery, we say consider giving a unit of RBCs, and the person gets it before anesthesia,” he says.

Dr. Daskal hopes next to conquer inappropriate use of fresh frozen plasma, most of which is used to reverse excessive warfarin anticoagulation. FFP carries major risks, such as transfusion-related acute lung injury. FFP also poses the risk of “volume overload, infections, and transfusion reactions—it’s not a benign process,” he says.

Dr. Daskal is working to achieve consensus at Albert Einstein that the blood bank will release FFP only for patients with a 1.9 INR or above who don’t have evidence of clinical bleeding.

Dr. Daskal also provides input for postoperative blood transfusion therapy. When a patient is bleeding after surgery, he goes to the unit to see if the bleeding exceeds a certain threshold. If he observes the patient has a chest tube with undiluted red blood that’s clotting, he suspects a surgical technical issue or “factor 14 deficiency,” as it’s called, in which case he talks to the surgeon.

To make blood management work, “you have to go to the floor and interact with the surgeons and see what’s happening. That’s how you develop trust,” Dr. Daskal says. He also uses “positive deviance,” whereby you identify the physicians who use less blood but have better outcomes. “You want to identify those resourceful individuals and make them the leaders.”

As a testament to the blood management program’s success, between three and four percent of the cases in Albert Einstein Medical Center’s liver transplant program are now bloodless. “The norm for liver transplants is three to 4.5 units per transplant on average,” Dr. Daskal says, and “only occasionally do we exceed seven units per transplant.” This he compares with the first transplant he was involved in in Houston in which the patient received 189 units, ultimately surviving but with “tremendous complications.”

Now, “I see these liver transplant patients who get minimal blood, and they are drinking coffee in their beds a few days after a liver transplant,” says Dr. Daskal. “They do as well as anyone.”

At Virginia Commonwealth University Medical Center, Richmond, the next step in the medical center’s two-year-old blood management program will be to explore ways to integrate bloodless medicine protocols.

Mary Jane Michael, RN, MSN, program coordinator for VCU’s blood management program, says VCU recognizes, of course, that patients need blood products in certain situations. “We aim to promote the judicious use of blood products for all our patients. We apply lessons learned from Jehovah’s Witnesses regarding blood conservation and alternatives to transfusion to all patients as a way to reduce their exposure to allogeneic blood.” This is especially important, she adds, to any patient who refuses a transfusion.

Susan Roseff, MD, VCU’s medical director of transfusion medicine, recalls a visit at VCU from an OB/GYN who had worked at a New Jersey hospital that practiced bloodless medicine. “He presented the case of a pregnant woman with placenta accreta having her third C-section who was at high risk for significant bleeding during delivery.” In such cases, the doctor would carefully manage the high-risk problem, maximize the patient’s own blood cell production before surgery by prescribing iron or erythropoietin, and use cell salvage along with other blood-sparing techniques during C-section, she says. But education is a big part of reducing blood use. Sometimes surgeons will order intraoperative FFP, cryoprecipitate, and platelets to treat a patient’s bleeding. “If they did some testing,” Dr. Roseff says, “they might see that the fibrinogen is very low and cryoprecipitate is the product of choice—or that only one component is necessary. Yes, FFP has fibrinogen, but it’s also much higher fluid volume.” For some transfusion practices, the jury is still out on what’s best, Dr. Roseff says.

“Oncology is a very, very big blood user, especially of platelets.” And oncologists appear to use different triggers for administering platelets to prevent bleeding. “One oncologist will use 50,000 as the cutoff and another 10,000 for patients receiving progenitor cell transplants,” she says. “Hopefully, a national study now underway will help determine the optimal platelet dose and transfusion frequency for these patients,” she says.

Of course, hospitals’ blood management efforts would be revolutionized if artificial blood could be substituted for donor blood.

Dan A. Waxman, MD, chief medical officer for the Indiana Blood Center, Indianapolis, says a number of trials over the last couple of years have studied artificial blood products or “oxygen carrying compounds,” as they are known.

While some of the products are promising, they have a short half- life and are at this point appropriate only for emergent situations, such as trauma or surgeries where there’s rapid blood loss, he says.

And “artificial blood comes with its own issues,” he cautions, noting that some patients who have received the products in clinical trials have had problems maintaining normal blood pressure.

Thus, “even with artificial blood,” Dr. Waxman says, “the same principles will apply as they do for human blood.” First, do no harm.

Karen Lusky is a writer in Brentwood, Tenn.

Strategic Healthcare Group is at www.bloodmanagement.com. Columbia Healthcare Analytics is at www.columbia-analytics.com.