In managing sepsis, the uphill battle continues

November 2007
Feature Story

Karen Lusky

Sepsis makes for a particularly perfect villain: It's notorious for taking healthy people to death's door in short order and has outwitted the best of medical minds trying to unravel its baffling trail of clues over the years.

Work is moving forward, however, in linking biomarkers to treatments for sepsis, though the advances come slowly and not without the usual stumbling blocks and controversy.

Currently, there are two FDA-approved biomarkers specifically for sepsis. One is Spectral Diagnostics' Endotoxin Activity Assay, or EAA, which measures the level of bioactive endotoxin in a patient's blood, with results in about 30 minutes. The FDA cleared the EAA in 2003 under a de novo classification to help identify patients at risk for progressing to severe sepsis within 24 hours of ICU admission.

The other is Berlin-based Brahms Diagnostics' procalcitonin, or PCT, test, a biomarker that can help flag bacterial infection in sepsis.

BioMérieux last month received FDA 510(k) clearance to market the Vidas Brahms PCT test in the United States to evaluate patients within the first 24 hours of their ICU stay. Brahms has a PCT test under FDA review now with potential applications for both the ICU and emergency department, says Jonas Leichtner, director of marketing for Brahms Diagnostics USA, Annapolis, Md.

Daniel Levine, PhD, director of the Iris & B. Gerald Cantor Clinical Research Laboratory at The Rogosin Institute and associate professor of biochemistry at Weill Medical College of Cornell University, New York, NY, thinks the EAA may help in determining the appropriate intervention in the ICU "or give you a way to figure out potentially who can go to a step-down unit, which is less costly."

The Endoxotin Activity Assay has been available for about four years, and more has become known about endotoxin and sepsis in that time. It may be that "the mediator of sepsis," says Spectral president and CEO Paul Walker, MD, PhD, is endotoxin translocated from gram-negative bacteria that normally inhabit the gastrointestinal tract.

That is, a person with sepsis can have endotoxin in his or her blood without having a gram-negative infection. Edward Abraham, MD, professor and chair of the Department of Medicine at the University of Alabama at Birmingham, says, "Gram-positive infection can lead to elevated endotoxin in the blood because as the person becomes sicker, his or her intestines become more permeable, releasing waves of endotoxin into the blood." Infection with viruses or even a parasite, such as malaria, can have the same effect.

"An endotoxin level in the setting of sepsis is very high," exceeding 0.6 EAA units, Dr. Walker notes. Using the EAA, Spectral has found that a normal value is about 0.26 EAA units. "The gray area is between 0.4 to 0.6 EAA units," he says.

Spectral is analyzing data now from a study conducted in the United States that looked at EAA levels in patients in the emergency department, though Dr. Walker says it's too early to release the results.

Thus far, the assay hasn't caught on in the United States, says Debra Foster, director of Spectral's sepsis program. The test, which doesn't have Medicare or private payer approval for reimbursement, she notes, costs less than half of what a blood culture costs.

Dr. Levine, who along with colleagues is doing research with endotoxin in an area outside of sepsis, says his experience in talking to physicians is that they are hesitant to integrate a new biomarker into their protocols unless it adds substantially to the tools they already have in place.

But Robert A. Balk, MD, director of pulmonary and critical care medicine at Rush University Medical Center, Chicago, attributes the assay's slow uptake in the United States to its nonspecificity. "It's quite sensitive to picking up endotoxin, but endotoxin may be present for a variety of reasons in critically ill people…" Periods of poor gut perfusion or altered mucosal barrier function, he notes, can result in translocation of the endotoxin or gram-negative organisms into the circulation. And that could lead to a false-positive EAA result indicating sepsis.

Spectral's Foster agrees that gut hypoperfusion can be a source of endotoxemia. Yet endotoxemia is an important risk factor for a patient having sepsis, she says, which is the essence of the company's FDA claim for the test. And as with all laboratory tests, Foster adds, the physician has to consider the result within the context of a particular patient and, if appropriate, intervene with early goal-directed therapy for sepsis.

Dr. Walker says Spectral has a preliminary hypothesis with data suggesting that adequate fluid therapy in sepsis may stop endotoxin from leaking out of the gut. "Adequate perfusion prevents organ dysfunction, and that is a very important component of treatment." An article in the November 2007 issue of Shock, he says, shows "fluctuating levels of endotoxin in the first three days in the ICU is very deleterious to the patient and is associated with multiple organ dysfunction" (Klein D, et al. Shock. 2007;28[5]:524-529).

If clinicians could identify patients with elevated or rising endotoxin levels, Dr. Walker says, they could treat them aggressively with fluids. "There may be a window of opportunity to get rid of endotoxin. Once the cascade gets downstream, it's hard to pull the person out."

The emergence of anti-endotoxin therapies offers new opportunities for combining the EAA with treatment. For example, in Japan and Europe, Dr. Walker says, there is an approved method for removing endotoxin from the blood: a dialysis process using a column, Toraymyxin (Toray Industries, Tokyo), coated with the antibiotic polymyxin B, which binds to endotoxin, stripping it out of the patient's blood.

David Klein, MD, coauthor of the November Shock article on endotoxin and a critical care medicine physician at the University of Toronto, notes there are anti-endotoxin therapies in clinical trials in the U.S., which, once approved, could provide a "pull effect" for use of the EAA.

Eisai Medical Research Inc., Ridgefield Park, NJ, is enrolling patients now in a phase three clinical trial to test an anti-endotoxin medication for severe sepsis. The investigational drug, eritoran tetrasodium, is believed to block the ability of endotoxin to activate toll-like receptor 4 (TLR4), says Mel Lynn, PhD, Eisai's associate vice president and global head of the sepsis and anti-infectives therapeutic area. "TLR4 is thought to play an important role in the course of severe sepsis, and when activated by endotoxin, can trigger a cascade of inflammatory responses."

The idea behind using procalcitonin—the only other FDA-cleared test for sepsis—is that it can help physicians identify which patients have clinically relevant bacterial infections so appropriate antibiotics can be initiated sooner.

An article published in Critical Care Medicine last year showed that adult patients with septic shock who received effective antimicrobial therapy within the first hour of documented hypotension had a survival rate of 79.9 percent. Each hour of delay in antimicrobial administration over the ensuing six hours was associated with an average decrease in survival of 7.6 percent (Kumar A, et al. Crit Care Med. 2006;34[6]:1589-1596).

Elevated PCT raises the index of suspicion that a patient has a bacterial infection causing sepsis. Herb Steward, executive vice president of BioMérieux, which is offering the PCT test on its Vidas and mini-Vidas immunoassay platforms with results available in 20 minutes, explains that PCT is found in the thyroid gland and is not normally present in the bloodstream of a healthy person. Multiple organs in the body release PCT when there is a systemic bacterial infection.

Typically in ICU patients, says Brahms Diagnostics' Leichtner, a cutoff of 0.5 ng/mL indicates a low risk of severe bacterial infection, whereas values of about 2.0 ng/mL are a clear indication of severe bacterial infection. In between, he says, is a gray zone.

In certain situations, Steward adds, PCT can rise without "any infectious aggression." Examples include trauma, burns, major surgery, and prolonged or severe cardiogenic shock. But the PCT levels usually return to normal rapidly. The physician can in such cases monitor the patient's PCT levels, and levels that don't recede may be an indication of bacterial infection, he says.

As a test for sepsis, PCT isn't without its detractors. A meta-analysis published in March by Benjamin Tang, MD, and colleagues at the University of Sydney, Nepean Hospital, in Australia, found that the test could not "reliably differentiate sepsis from other non-infectious causes of systemic inflammatory response syndrome in critically ill adult patients" (Tang B, et al. Lancet Infect Dis. 2007;7[3]:210-217).

In a response to the study's findings published in the same journal, Mueller, et al. (2007;7[8]:498-499) point out that PCT as a marker for serious bacterial infection has been studied in both observational and intervention trial designs. They write, "A gold standard to differentiate infectious from non-infectious causes in patients with SIRS [systemic inflammatory response syndrome] does not exist, and therefore all observational studies are prone to a potential bias."

Some studies use the physician's decision to order blood cultures or to prescribe antibiotics as the definition for bacterial infection, Leichtner says. Other studies rely on infectious disease specialists to identify bacterial infection or look back retrospectively at laboratory data and other findings. Thus, he says, such studies may not be comparing apples to apples.

By contrast, intervention trials involve use of PCT to guide which patients require antimicrobial therapy. The thinking is that patients with a low PCT do not have a serious bacterial infection and can go without antibiotics. "For PCT," Leichtner says, "there are several randomized controlled trials that have successfully evaluated this concept in patients with low respiratory tract infections."

Rush University's Dr. Balk believes that combining PCT with polymerase chain reaction assays and clinical assessment may more quickly identify patients with sepsis due to bacterial infection, including gram-negative infection. "You would then initiate early antibiotic therapy based on the likely organism(s) and your knowledge of the antibiogram" of the institution, community, or both, he says.

BioMérieux sees PCT along with molecular infectious disease testing as a potentially winning combination for sepsis. At this year's American Society for Microbiology meeting, the company announced the exclusive distribution agreement with AdvanDx for PNA FISH, molecular-based tests that make it possible for labs to more rapidly identify infectious organisms directly from blood cultures.

Margie Morgan, PhD, a microbiologist at Cedars-Sinai Medical Center, Los Angeles, says PNA FISH, which "is easier to perform and less expensive than PCR," is a staining technique performed on positive blood cultures that can in 2.5 hours identify Staphylococcus aureus, coagulase-negative staph, enterococcus, Candida albicans, and Candida glabrata.

BioMérieux signed an exclusive deal with Cepheid to develop a line of sepsis test products on Cepheid's GeneXpert platform. The GeneXpert system's real-time PCR amplifies and detects DNA sequences associated with a microorganism. The test menu is scheduled to be launched in Europe in 2010 and in the United States in 2011. Cepheid will manufacture the assays, and BioMérieux will distribute them on an exclusive worldwide basis, Steward says. The product menu consists of bacterial and fungal identification assays and a series of genetic markers for antibiotic resistance.

So far, there is only one specific sepsis treatment, Eli Lilly's Xigris (activated protein C or drotrecogin alfa), approved by the FDA in 2001 for patients with severe sepsis at high risk for death. Lilly has a study underway, RESPOND, that involves measuring serial protein C levels in patients taking Xigris to evaluate their therapeutic response to the drug.

Lilly is conducting the study with Inverness-owned Biosite Inc. in an effort to develop a point-of-care protein C test that could be used to adjust the dosage and duration of Xigris treatment based on a patient's protein C levels, says Mark Williams, MD, medical director for Xigris global brand development at Eli Lilly and Company, Indianapolis.

Sepsis patients with a lower protein C level have been shown to have a higher risk of death. And, says John Brandt, MD, clinical research physician/pathologist at Eli Lilly, "Administration of Xigris, a recombinant activated protein C, has been associated with a more rapid increase in the levels of endogenous protein C." Sepsis patients whose endogenous protein C levels return to normal appear to have higher rates of survival.

The hope is that using protein C testing for patients taking Xigris will improve the benefit-risk profile for the medication, which can cause bleeding.

The PROWESS study leading to FDA approval of the drug showed that the medication significantly reduced mortality for people with severe sepsis. But a followup clinical trial, the ADDRESS study, investigating the use of Xigris in patients with less severe sepsis, was not favorable, Dr. Williams says. It showed that Xigris failed to benefit the lower-risk group of patients and that the incidence of bleeding was almost identical to that seen in the PROWESS study.

Responding to a request for more data on Xigris from the European equivalent of the FDA, Lilly has a global study scheduled to start in March 2008 called PROWESS-SHOCK. That trial will involve about 1,500 patients over about two years to see if Xigris reduces mortality in septic shock, Dr. Williams says. The primary end point is 28-day mortality from all causes, which is the classic end point for a sepsis trial.

Thanks to animal model studies, Lilly now has a better grasp on Xigris' mechanism of action than it had in the initial PROWESS trial. "We know that Xigris has an anticoagulant effect, but now we believe that the drug's main effect is as an anti-inflammatory agent," Dr. Williams says. The drug may reduce inflammation "by binding to white blood cells and decreasing their rolling and adhesion to the endothelium and through specific receptor signaling pathways."

The University of Alabama's Dr. Abraham says that because Xigris appears to have multiple mechanisms of action, the Xigris trials may help physicians and scientists better understand the underlying pathology of sepsis.

Work continues on identifying sepsis biomarkers that signal impending organ dysfunction in time to stave it off.

One emerging biomarker in the early stages of development, the protein neutrophil gelatinase-associated lipocalin, or NGAL, shows promise for detecting acute renal failure associated with sepsis perhaps in time to turn it around, says Prasad Devarajan, MD. He is the Louise M. Williams Endowed Chair and professor of pediatrics and director of nephrology and hypertension at Cincinnati Children's Hospital Medical Center.

"In sepsis, it isn't usually the bug that kills you—it's the body's response to the bug that creates both pro- and anti-inflammatory responses that ultimately kills the person," says Dr. Devarajan. And part of that response is development of acute renal failure.

"If you take all the causes of acute renal failure, sepsis leads the list as the No. 1 cause," he says. Of the 800,000 people who develop sepsis in the U.S. annually, about half develop acute renal failure. And 25 percent of those require dialysis, a group that has a mortality rate exceeding 80 percent.

Yet a timely diagnosis of acute renal failure in sepsis is hard to make, just as it is for other forms of acute renal failure, because by the time physicians use the "gold standard—an increase in creatinine—it's far too late to intervene," Dr. Devarajan says.

"We have published our first study showing that in patients who have sepsis, urine NGAL can predict ARF one to two days before creatinine goes up. We have a paper in revision now showing that plasma NGAL does the same thing."

What might you do differently for a patient if you knew he or she had elevated NGAL with sepsis? "That's the patient you'd want to monitor very carefully in the ICU rather than putting him on a regular floor," Dr. Devarajan says. "You'd have to be very careful about maintaining the patient's kidney perfusion and avoiding nephrotoxic agents, such as IV contrast for imaging or certain antibiotics. You need to be informed to do no further harm."

Even in patients who do not develop full-blown acute renal failure, he continues, a large volume of literature suggests physicians should intervene with renal replacement therapy, such as hemofiltration techniques that remove endotoxins and cytokines and related products in the blood that add to the systemic disease.

In addition, he says, experimental therapies for acute renal failure have been shown to work "beautifully" in animal models. Yet researchers have spent millions trying the agents in humans without success. "We think the reason they may not work is that they are administered far too late—after creatinine goes up," Dr. Devarajan says. Having an early biomarker for acute renal failure, such as NGAL, allows researchers to redesign the studies and test a long list of 20 to 30 medicinal treatments earlier to see if they improve outcomes.

The majority of studies on NGAL thus far, however, have been single center studies that have been reasonably well done, but they're small studies nevertheless, Dr. Devarajan cautions. So more work needs to be done.

The University of Toronto's Dr. Klein observes that NGAL seems to have a role as an early marker of renal dysfunction. And "hopefully in the future, that can be tied to therapies to prevent or treat early renal dysfunction potentially in sepsis."

The more that unique populations of sepsis patients can be identified to target with specific, directed therapies, the better. In doing so, Dr. Klein says, "perhaps we can undo the failures of the last 20 years where billions have been spent on sepsis therapies that have failed, probably in part because they were applied in too wide of a clinical context so the wrong patients received them."