stanford medicine


Melinda Gates on family matters - A conversation about contraception

Special Report

Against the flow

What’s behind the decline in blood transfusions?

One day in 2011, an ambulance pulled up to the Stanford emergency room and paramedics unloaded a man in his 30s who had crashed his motorcycle. He was in critical condition: Tests showed dangerously low blood pressure, indicating that around 40 percent of his blood was lost. And an ultrasound revealed that the blood was collecting in his belly, suggesting that one or more of his abdominal organs was the source of the blood loss.

Paul Maggio, MD, a trauma surgeon and co-director of critical care medicine at Stanford Hospital & Clinics, sped the patient into the operating room. But he made sure that the technicians prepping his operating room took the time to set up one key piece of equipment, called an intraoperative cell salvage device, which is now commonly used in trauma cases. As the patient lay on the operating table and Maggio made the first cuts into his abdomen, suction devices slurped up the loose blood, directing it away from the surgery site through tubes. But instead of leading to a container bound for disposal, the tubes led to the salvage device. 

The ATM-sized machine spun the blood to separate its components, cleaned it of any debris that had been suctioned up from the abdomen and sent it back out into fresh bags. From there, the blood was shunted right back to the patient’s body, through intravenous tubes poking into his veins. The cell salvage device has been around for decades, but only recently has evidence emerged that autotransfusion — giving patients their own blood instead of blood from donors — leads to better surgery outcomes. As a result, the use of the machines has gone from extremely rare to commonplace. Today, hospitals that have the machines use them in many scheduled abdominal and heart surgeries and routinely in trauma cases involving massive bleeding.

“Autotransfusing this patient spared him from getting more banked donor blood and from all the risks associated with it,” says Maggio of the motorcycle crash victim. He turned out to have an injury to his spleen, which Maggio repaired. In all, around 2 liters of blood were collected from the patient’s abdomen, processed through the salvage device, and transfused back into his body. 

Blood transfusions involve routing a needle into one of a patient’s veins — most often in an arm — and attaching a thin tube to the needle. Blood flows through the tube directly into the patient’s blood vessels. Ten years ago, a patient like Maggio’s would most likely have had a transfusion of blood donated by volunteers at the Stanford Blood Center. But over the past decade, a growing body of research has revealed that in hospitals around the world, donated blood is used more often, and in larger quantities, than is needed to help patients — both in operating rooms and hospital wards.

Some of the research has been conducted by physicians working with patients who refuse donated blood on religious grounds; other findings have come from the front lines of the war in Afghanistan, where blood is hard to transport; and some studies have been inspired simply by the rising cost of blood and a desire to save resources. Some findings are new, and others, like studies by Stanford’s Tim Goodnough, MD, a hematologist and the director of transfusion services, are years old but only recently being noticed. The takeaway message from all is the same: While blood is precious and continues to save lives, its use can be minimized and fine-tuned to optimize patients’ health and reduce costs.

The American Medical Association brought attention to the subject last fall at its national summit on the overuse of five medical treatments. Blood transfusions were on the list (along with heart stents, ear tubes, antibiotics and inducing birth in pregnant women). 

“From the clinical standpoint, I’m not really thinking about resources or cost,” says Maggio, who’s also an assistant professor of surgery. “I’m thinking about giving the patient the best care.” Donated blood carries risks, albeit very slight, of infection and setting off an immune reaction. But research is also showing that even when these drastic outcomes are avoided, there’s something else about donated blood — which scientists don’t fully understand — that could slow recovery time or increase complications.

While autotransfusion for trauma patients is growing, and guidelines for blood transfusions are changing in response to this new research, altering the protocols that doctors have been using for so many years is a slow process.

Changing the routine

At Stanford, it took an innovative new program that used alerts on doctors’ computer systems to enforce fewer blood transfusions. But the push paid off: Blood use in the operating rooms, emergency rooms and hospital wards of both Stanford and the Lucile Packard Children’s Hospital has declined by 10 percent in just a few years. At Packard Children’s alone, 460 transfusions and $165,000 were saved in one year, according to a pilot study conducted Feb. 1, 2009, through Jan. 31, 2010.

‘There’s this idea ingrained in the culture of medicine that people will die if they don’t have a certain level of blood, that blood is the ultimate lifesaver.’
Patricia Ford, MD, Founder and Director of Pennsylvannia Hospitalís Center for Bloodless Medicine and Surgery at Penn Medicine

“I think we’re probably still giving too much blood in some of these situations,” says Maggio. “But we hope that physicians are becoming better informed about when to give blood.”

People most often need blood transfusions when they’re in one of three situations: They lose blood from a major surgery that’s been scheduled for weeks or months; they lose blood in a way that their body won’t be able to replace, such as a blood cancer that shuts down the body’s ability to make blood cells; or they lose blood during a more sudden trauma — either an external wound or internal bleeding.

“For that first group of patients, scheduled for elective surgery, if you can plan ahead, you should be able to avoid using blood,” says Goodnough, a professor of pathology and of medicine. In those patients, drugs can boost a patient’s own blood production ahead of surgery, blood can be collected from a patient ahead of time to re-infuse later, precautions can be taken to prevent sudden blood loss, or autotransfusion machines like the cell salvage device can be set up. “Where we still need a national blood inventory is for patients who can’t plan ahead,” says Goodnough.

In the cases where physicians continue to give blood when it might not be needed, it’s often because they can’t imagine not doing everything they can to help a patient — and blood has always been viewed as having far more benefits than risks in almost any population of patients. But now, that risk-benefit analysis is changing.

“There’s this idea ingrained in the culture of medicine that people will die if they don’t have a certain level of blood, that blood is the ultimate lifesaver,” says Patricia Ford, MD, founder and director of Pennsylvania Hospital’s Center for Bloodless Medicine and Surgery at Penn Medicine. “And that’s true in some specific situations, but for most patients in most situations it’s just not true.” Ford’s center is one of the oldest and largest in the country that specializes in treating patients without donated blood; dozens of others have been created over the past decades but mostly at a smaller scale.

Going bloodless

Every year, Ford treats or operates on around 700 Jehovah’s Witnesses, whose religion prohibits transfusions of blood that is not one’s own. Since 1996, she has been fine-tuning ways to give these patients the best care as well as ways to apply these techniques to the broader population.

“Many physicians I talked to at the beginning had this misperception that a lot of patients just can’t survive without receiving blood,” says Ford. “I may have even thought that myself to some degree. But what I rapidly learned was you can care for these patients by just applying some easy strategies.”

In fact, a study published in August 2012 by researchers at the Cleveland Clinic concluded that Jehovah’s Witness patients recovered better from heart surgery than patients who received blood transfusions. It’s the longest study conducted on such patients — the researchers followed them for up to 20 years. The Jehovah’s Witness patients had higher five-year survival rates, fewer heart attacks following the surgery and fewer complications including sepsis and renal failure. The better outcomes might not have been due to the absence of transfusions but to differences in care received — the patients were more likely to be treated for low blood levels before surgery by receiving iron supplements and vitamins, and every patient’s surgery included use of an intraoperative cell salvage device. The findings suggest that these methods employed for bloodless surgeries could help patients beyond the Jehovah’s Witness community. 

At Pennsylvania Hospital, Ford has discovered that, for scheduled surgeries, one of the best ways to avoid the need for blood transfusions is to test patients’ levels of hemoglobin — the protein in red blood cells that carries oxygen — well before their surgery. If the levels are low, then the patient can take vitamin K and iron supplements, which help the body produce more blood cells and help red blood cells more efficiently carry oxygen throughout the body. The practice of testing for low red blood cell levels, or anemia, is now beginning to spread from specialized clinics like Ford’s to other hospitals around the country. 

“Testing for anemia was just not on people’s radar screens, because they knew that they could always give the patient blood,” says Ford. Now, many doctors consider testing a patient’s blood cell levels just as important as testing their heart and lung health before surgery. This shift is supported by studies such as an October 2012 analysis in the Annals of Thoracic Surgery of the outcomes of more than 17,000 heart surgeries, which found an increase in stroke, death during surgery and death after surgery when patients were anemic before surgery.

At Stanford, standard pre-surgery tests include blood counts for patients who are expected to lose large amounts of blood, says Goodnough. If anemia is suggested by the results, clinicians aim to manage the condition before surgery.

At Penn, Ford also emphasizes the conservation of blood during surgery, often by using an intraoperative cell salvage device. Patients can also donate blood in the weeks leading up to a scheduled surgery and their own saved blood — called an autologous donation — can be used for a transfusion if necessary. In the 1980s, Goodnough studied the usefulness of autologous donations in different patient population groups and pushed for its broader usage. It’s now considered a mainstream way of reducing the need for donated blood. “It sounds like a mundane concept now, but it was quite progressive when we first started looking at it,” says Goodnough.

Among Ford’s lessons with the Jehovah’s Witnesses, she says that perhaps her most important has been that there’s no magic hemoglobin number that tells doctors when a patient will start exhibiting signs of anemia. Typically, doctors consider hemoglobin above 12 to be normal, and hemoglobin below 7 or 8 to indicate the need for a blood transfusion. But Ford and a growing number of other doctors think those numbers could be pushed down further, a change that would require new studies for many to adapt.

“It’s not unusual for me to see a patient who has a hemoglobin of 5 and they look as healthy as anyone walking down the street,” says Ford. Of course, there also can be patients who become sick with much higher hemoglobin levels, but Ford would like to see more doctors treating blood levels based on symptoms, not a number. Goodnough agrees: “It’s really hard to demonstrate at what level of hemoglobin a transfusion will help a patient,” he says. “And we’re increasingly seeing that for most patients, hemoglobin has to be exceptionally low to have effects.” But it depends more on the patient’s health and risk factors, he says. There’s no one-size-fits-all solution.

Applying lessons

Beyond learning from Jehovah’s Witness patients who receive no donated blood during surgeries, the past decade has seen the first controlled trials in the broader population to test whether limiting blood transfusions — though not eliminating them entirely — affects outcomes. The first large trial to test whether allowing patients to have lower levels of hemoglobin was harmful was called Transfusion Requirements in Critical Care and the results were published in 1998. Before then, transfusing any critically ill patient with hemoglobin levels under 10 grams per deciliter was considered appropriate — even necessary — treatment. But the trial looked at the outcome of 800 patients in intensive care units and found that there was no difference in patients’ health over 30 days if the transfusion trigger was 7 instead of 10. Moreover, the amount of blood used by an ICU was halved when the trigger was lowered. But the findings applied only to a specific patient population.

“What was happening in the community is that everyone was focusing on this one initial study but it was done on ICU patients only,” says Jeffrey Carson, MD, of Robert Wood Johnson Medical School. “We figured those results really needed to be replicated in other patient populations.”

So Carson and his colleagues launched a new study, following more than 2,000 surgical patients who had hemoglobin levels less than 10 after hip surgery and had accompanying cardiovascular disease, making them a higher-risk population. Once again, they found that neither death nor complication rates increased with a more restrictive transfusion strategy. The results were published in the New England Journal of Medicine in 2011. In Carson’s trial, one-third the amount of blood was used in hospitals when the transfusion trigger was dropped from 10 to 7.

But more studies are still needed. “The question is, how low can you go?” Carson says. If 6.5, or even 6, is as good a trigger as 7, blood is still being transfused unnecessarily.

At Stanford and Lucile Packard Children’s Hospital, physicians including Goodnough spearheaded that next step — enforcing less blood use — by creating a unique method of nudging doctors to change their ways.

An analysis at Stanford revealed that more than two-thirds of patients with hemoglobin levels over 8 grams per deciliter were receiving blood transfusions. “It led us to believe we were heavily overutilizing blood,” says Goodnough. 

Now, if a Stanford doctor tries to order a blood transfusion for a patient with hemoglobin levels over 8, a pop-up alert appears on the computer, reminding the doctor of the latest guidelines on when to transfuse blood (the AABB published guidelines in July 2012 recommending transfusion only below a hemoglobin level of 7 for stable patients), and asks questions about the doctor’s reason for ordering the transfusion. The doctor ends up cancelling the order 40 percent of the time.

“The system helped people think twice,” says Goodnough.

For those 60 percent of incidents in which doctors continue to give blood to patients with a hemoglobin level above 8, the reasoning is occasionally based on a patient’s symptoms — if they don’t have stable vital signs, for instance, there is evidence that a transfusion could help them. But in many cases, it’s simply physician preference; doctors want to stick to the protocols they have been using for years, even though they may be outdated.

Nationwide there’s been a dramatic decrease in blood usage, says Susan Galel, MD, director of clinical operations at the Stanford Blood Center. Some of it is attributed to the economic downturn combined with the rising cost of blood — hospitals can save money by buying less blood — and some of it to more effective blood management and efforts like Stanford’s computer system.

In a May 2012 Anesthesiology article, Goodnough quoted Richard Benjamin, the chief medical officer for the American Red Cross, as stating in a personal communication that national blood usage had declined 3 percent each of the previous two years, for a cumulative decline of almost 7 percent. The details are not public at this time.

Whatever the motivation for the decrease in blood use, and whatever the pace of the decline, the outcome of using less blood is conserved resources, saved money and — based on the recent studies — improved patient health.

Cases of trauma

In some emergency situations, though, patients will always need blood transfusions. Despite years of research, no safe substitutes for real human blood have been developed that can help patients in cases of blood loss. Substitutes that showed promise in the 1990s failed in clinical trials after leading to increases in mortality rates and adverse outcomes such as heart attacks, and exciting early prospects have led to disappointment time after time. As recently as the early 2000s, the compound PolyHeme was generating excitement and being tested in phase-3 clinical trials. But the trials were halted in 2006, citing an increase in patient deaths when relying on the blood substitute, and temporarily, at least, closing the door on artificial blood. So today, doctors continue to depend on donated blood from blood banks when they’re dealing with major emergency blood loss. 

Now if a Stanford doctor tries to order a blood transfusion for a patient with hemoglobin levels over 8, a pop-up alert appears on the computer. “The system helps people think twice.”
Tim Goodnough, MD, Stanford Hematologist, Director of Transfusion services and professor of Pathology and of Medicine

Even in these trauma situations, however, where it’s clear that the need for blood is not going away any time soon, doctors are constantly researching the best ways to use blood products to help patients survive: When should blood be given? How much? What mixture of blood components should a transfusion consist of?

“What we’re doing is coming up with massive transfusion protocols to help physicians administer blood products in the right ratios,” says critical-care specialist Maggio.

While most blood transfusions — those that are discussed above — consist of red blood cells, there are other components of blood that may need to be replaced in patients suffering massive blood loss. So the mixture typically given to trauma patients has three ingredients: red blood cells, plasma (the liquid part of blood that contains not only water, but sugars, proteins, fats and salts) and platelets (fragments of cells that help blood clot).

And for patients who need these fine-tuned mixtures of blood, it’s not always the case that less blood is better. Instead, it’s a matter of recognizing when a patient does need massive amounts of blood and determining how it’s best administered. In the past, Maggio says, a doctor’s first priority when seeing a patient who had lost a large amount of blood was to provide a solution mimicking the fluid component of blood — not the blood cells. But recent studies have shown that for patients requiring large volume resuscitation, administering blood cells should be a higher priority than fluids. So Stanford has a new system to make that possible.

When a patient with major blood loss is seen in the emergency room, doctors can immediately activate the massive transfusion protocol, says Maggio. “A cooler of blood comes down to the ER containing six units of packed red blood cells, four units of fresh frozen plasma and one pack of platelets.” This ratio of red blood cells, plasma and platelets approximates whole blood. Data from the military suggests that survival is improved in patients who received the components at ratios similar to blood, but studies have yet to show whether switching to whole blood, rather than mixing the components, shifts that benefit. “A lot of these changes to transfusion practices in trauma are actually based on military data coming out of Afghanistan and Iraq,” says Maggio. “And we’re now applying it to the civilian population.”

At Stanford, trauma doctors are also adopting blood reuse protocols such as those that Ford uses at Penn. For major trauma patients, like the motorcycle crash victim he saw last year, the cell salvage device is critical, Maggio says. And it’s routinely set up for every case of major trauma at Stanford. “For cases where I suspect there may be a large amount of blood loss, I also call ahead and make sure it’s set up,” Maggio explains. Once surgery has begun, it’s too late to set up the device, since the majority of blood is cleaned out of the surgery site immediately after the first incision, he says.

And for many trauma patients, and others who need a transfusion in an emergency, it’s still true that blood is a lifesaver. The more than 50,000 blood donations through Stanford Blood Center each year are key to saving many lives.  

But for patients whose hemoglobin levels are borderline and appear healthy, doctors are thinking twice about whether the cost and risks of blood transfusions are worth it. “Physicians don’t understand well the benefits of blood transfusion. We’re pretty sure it saves lives but that’s never been demonstrated for the FDA,” says Goodnough. “There’s never been a prospective trial demonstrating that blood saves lives.”

Such a trial will likely never happen, at least in the full-blown sense, but the studies of Jehovah’s Witnesses and the lack of ill effects from lowering the transfusion trigger point have made it clear that in some cases the human body can recover from low blood levels by relying on its own, natural mechanisms for blood replacement.

So when Goodnough sees patients in the hospital wards, he does what every doctor is trained to do: judge patients’ health through not a single number or test, but by a combination of factors, including patients’ own reports of how they feel. Hemoglobin levels are only one part of the puzzle as to how someone’s body is operating, and transfusions of blood are only one possible course of action if symptoms do suggest low blood levels.


E-mail Sarah C.P. Williams






©2013 Stanford University  |  Terms of Use  |  About Us