Home Feature A review of the evidence to treat sepsis and anemia

A review of the evidence to treat sepsis and anemia

December 2005

Published in the December 2005 issue of Today’s Hospitalist

With more than 750,000 new cases a year in the United States and a mortality rate of up to 50 percent, sepsis is a serious problem. The condition kills more than 1,400 Americans a day, making it the leading cause of mortality in the ICU.

But sepsis is often difficult to diagnose, at least initially, because patients display signs that can be interpreted as other common conditions. And even once a diagnosis has been made, treatment can be challenging.

There are several drug therapies to treat sepsis, but not all of them are as effective as you may think. In addition, recent data have shown that some commonly accepted therapies may do more harm than good.

Here’s a review of the evidence presented by William Janssen, MD, instructor of medicine at the University of Colorado, on two strategies: drotecogin alfa, also known as activated protein C, and packed red blood cell transfusions. Dr. Janssen spoke at the Fall 2005 Hospitalist CME Series, which was held in cooperation with Today’s Hospitalist magazine.

Weighing the pros and cons of activated protein C

Because sepsis often leads to problems like disseminated intravascular coagulation (DIC), Dr. Janssen said, a relatively new drug is generating interest among physicians who work in the ICU.

Drotecogin alfa, which is also known as activated protein C or Xigris, was studied in the PROWESS trial, the results of which were published in a 2001 issue of the New England Journal of Medicine.

Researchers identified nearly 2,500 patients who developed severe sepsis or septic shock and gave them either a placebo or activated protein C continuously for 96 hours. The study found a 6 percent drop in relative mortality among patients who received the drug.

Dr. Janssen said the drug was so successful in reducing mortality, in fact, that researchers stopped the trial early. Investigators decided it was unethical to deny the drug to patients who had been receiving placebo.

Significant bleeding risks

Despite the trial’s success, Dr. Janssen noted that there are several caveats to using activated protein C. There was a long list of patients who were excluded from the trial and presumably are not candidates for the therapy. They include children, as well as patients who are obese, pregnant, breastfeeding and have nutropenia.

But perhaps even more importantly, patients at risk of bleeding need to be carefully weeded out. The study, for example, excluded patients with chronic renal failure and chronic liver failure, patients taking anticoagulants, and patients with thrombocytopenia. It also excluded patients who had recently experienced bleeding from severe trauma, had a history of severe CNS lesions, experienced GI bleeding within the previous six weeks, or undergone major surgery in the previous 12 hours.

Even after investigators excluded patients at high risk of bleeding, they still found that bleeding was a serious problem. “During the infusion period,” Dr. Janssen said, “patients who received activated protein C were twice as likely to have a serious bleeding event as those who received placebo. This rate was 2 percent, which is comparable to the drugs that we use for myocardial infarction, such as IIB/IIIA inhibitors.”

While the biggest problems came from intra-abdominal and intrathoracic bleeding, intracerebral bleeding rates were relatively low.

Organ failure and APACHE scores

When researchers further studied activated protein C’s effect relative to organ failure and APACHE scores, the drug’s value became more clear.

“When you look at organ failure,” Dr. Janssen explained, “you see that the patients who had only one organ failure did not appear to have a benefit from the drug, whereas patients who had multiple organ failure appeared to derive the most benefit.” He noted that in Europe, the drug is used in patients who have severe sepsis and who have two or more organ failures.

When investigators looked at the drug’s effect relative to patients’ APACHE scores, they found another significant result: When the patient’s APACHE score was 24 or less, the drug produced relatively little benefit. When the APACHE score was greater than 25, however, the drug was associated with an improvement in mortality.

Dr. Janssen said that another way to assess activated protein C is to look at it in terms of number needed to treat. When the drug is given to patients with APACHE scores of less than 25, the number needed to treat is only eight. That number, he added, is low when compared to other interventions used in the hospital.

Practice recommendations

How can you balance the potential mortality-reducing benefits of activated protein C with downsides like bleeding?

For one, Dr. Janssen said, pay attention to platelet counts. “Bleeding is a major concern,” he explained, “but the risk is highest when the platelet count is less than 30,000.” And because patients with DIC benefi t most, he suggested checking these patients’ platelet counts fi rst. “If their platelet count is greater than 30,000,” Dr. Janssen said, “this drug is recommended for use.”

Once these patients’ platelet counts stabilize, he added, don’t assume it’s safe to stop activated protein C. “If you start the drug and their platelet count drops,” Dr. Janssen said, “many experts would advocate transfusing platelets rather than stopping the drug.”

Finally, to minimize bleeding risks, avoid giving the drug before and after surgery. “If you’re going to do an invasive procedure or the patient is going to have surgery,” he added, “the infusion should be stopped two hours prior. You can restart it 12 hours after surgery or two hours after a less invasive procedure.”

At his 35-bed ICU, which is typically full, Dr. Janssen said he uses activated protein C two or three times a month. He acknowledged that activated protein C is expensive and costs $6,800 for a four-day course. He also noted that several investigators have conducted a cost-benefit analysis that estimated the drug costs about $20,000 for every year of life saved.

That figure is relatively high, he said, because the mortality rate for sepsis is so high. While studies typically focus on 28-day mortality, he said that sepsis survivors have a 12-month mortality of only 12 percent after discharge. For patients who have an APACHE score of greater than 25, that rate climbs to 17 percent. Once patients survive a year, their mortality rates go down sharply.

“We really need more interventions that can improve long term survival in these patients,” Dr. Janssen said.

How blood transfusions can hurt your critically ill patients

You’re probably well-versed in the pros and cons of giving critically ill patients packed red blood cell transfusions, but did you know that the strategy can often backfire and raise patient mortality?

Dr. Janssen explained that between 30 percent and 50 percent of patients admitted to the ICU will require a transfusion at some point during their stay, and many will need almost fi ve units of blood. But recent research shows that for some patients, at least, transfusions can do more harm than good.

Dr. Janssen said that while most physicians realize that transfusion can transmit infections and cause acute lung injury, it can also impair immune function.

“We’re starting to recognize more and more immune modulation,” he explained. “We know that patients who have trauma and receive blood transfusions are at greater risk of sepsis.”

Dr. Janssen presented the results of a study published in a 2002 issue of the Journal of the American Medical Association (JAMA) by Vincent et al. that found an increased mortality rate among patients who received a transfusion in the ICU. Those who received a transfusion had a mortality rate of 19 percent compared to 10 percent of patients who didn’t undergo the procedure.

Dr. Janssen said the results raise an obvious question: Wasn’t mortality greater among patients who received a transfusion because they were sicker? To answer that question, researchers created a risk score by looking at factors like age, diagnosis and the presence of chronic diseases. “They found that patients who received a transfusion still had a higher mortality compared to those who did not receive a transfusion,” he said, “even when all things were considered.”

Hemoglobin as a marker

Another study, the Transfusion Requirements in Critical Care (TRICC) trial, took another look at the issue by examining nearly 900 patients who developed anemia during their stay in the ICU.

While researchers excluded patients with chronic anemia and an admission hemoglobin of less than 9 or active bleeding, they did not exclude cardiac patients.

The study, which was published in the New England Journal of Medicine in 1999, then randomized patients to two groups. When treating patients in the fi rst group, physicians used a “liberal” transfusion strategy with a target hemoglobin of between 10 and 12. Physicians ordered a transfusion when patients reached a hemoglobin of 9.

When treating patients in the second group, physicians used a “conservative” strategy with a target hemoglobin of between 7 and 9. Patients in this group received a transfusion when they reached a hemoglobin of 7.

When mortality rates for all patients were reviewed, researchers found that the in-hospital mortality rate was better for patients managed conservatively (28.1 percent vs. 22.2 percent). They found relatively little difference in 30-day mortality rates for all patients.

When the study looked at relatively healthy patients ” individuals with an APACHE II score of less than 20, for example “signifi cant differences appeared. Less acutely ill patients in the conservative strategy group had a lower 30-day mortality mortality rate (8.7 percent) compared to patients in the liberal strategy group (16.1 percent).

“These are young, healthy patients,” Dr. Janssen said. “It’s pretty obvious that those patients can tolerate a lower hemoglobin. It’s safe to let those patients drift down to 7 before transfusion.”

To examine the effects of transfusion on a less healthy group of patients, researchers examined patients with clinically significant ischemic heart disease. They found that giving these patients a transfusion didn’t affect mortality rates between the two groups.

“It appears to be safe to use a lower transfusion trigger in patients with ischemic cardiac disease,” Dr. Janssen concluded.

Practice recommendations

Dr. Janssen acknowledged that the evidence from these trials may appear to be at odds with the notion that transfusions can help volume resuscitate patients in the early stages of sepsis. Earlier in his presentation, in fact, Dr. Janssen emphasized the importance of transfusions in early goal-directed therapy for sepsis patients.

How do you reconcile the different data on the value of transfusions for critically ill patients? Dr. Janssen recommended the following approach: If your patient is septic and has a low Sv02, give a transfusion to bring the patient’s hematocrit above 30.

“I would continue to do that in sepsis when the patient is unstable and when they have poor tissue delivery of oxygen,” he explained. “Once they become stable and their blood pressure is normal, however, use a conservative infusion strategy.”

“This is a difficult area,” he acknowledged. “Many of us have been taught to transfuse patients who have a hemoglobin of less than 8 or patients with severe coronary disease who are less than 10. I’m not necessarily saying that you should have a threshold of 7 for all patients. What I am saying, however, is that you should treat the patient, not the lab value.”

He also said that using erythropoietin to rescue critically ill patients is only marginally effective. Data from a trial published in 2002 in JAMA by Corwin et al. found that patients who received erythropoietin required fewer transfusions than patients who received placebo, but that a benefit was not reached until about eight days. In addition, there was no change in all-cause mortality.

“Data don’t support the routine use of erythropoietin in the ICU,” Dr. Janssen said. “You can consider erythropoietin plus iron for patients who will be in the ICU more than seven days, and probably those who are going to go to a chronic care facility, but there is probably no acute role for erythropoietin at this time in the ICU.”

Edward Doyle is Editor of Today’s Hospitalist.

For more on the early detection and treatment of sepsis, see “Time is tissue: Why emerging evidence on sepsis urges physicians to watch the clock,” in the August issue of Today’s Hospitalist online at www.todayshospitalist.com.

Do transfusions really deliver more oxygen to critically ill patients?

One of the benefits of transfusions is thought to be increased oxygen-carrying capacity, but does this improve oxygen consumption? The answer is “maybe.”

Blood stored at your local blood bank has a shelf life of about 30 days. So when an order for blood comes in, blood banks typically move out the oldest product first to avoid wasting any of the precious red stuff.

As blood ages, explained William Janssen, MD, instructor of medicine at the University of Colorado, the levels of 2,3 DPG, one of the factors related to oxygen dissociation, tend to drop.

“As these levels drop,” Dr. Janssen said, “the erythrocyte can pick up oxygen more avidly, but it has difficulty delivering or unloading oxygen. So when you transfuse blood, you have plenty of red cells carrying oxygen, but they might not be able to unload it to the tissues.”