Feature Story

Chasing after sepsis markers, test by test

December 2004
Sue Parham

Three years ago, an informal evaluation by Patrick St. Louis, PhD, of the Department of Clinical Biochemistry, and his clinical colleagues at Sainte-Justine Hospital, Montreal, suggested that 80 to 90 percent of patients in the neonatal intensive-care unit suspected of having sepsis were treated with antibiotics without having an infection. This was substantiated by a national study based on mean NICU admissions, established at approximately 650 per participating hospital, showing that 50 out of 650 had at least one episode of confirmed infection but 480 out of 650 received antibiotics.

"Given this data, it appears that a huge number of patients are receiving antibiotics inappropriately," says Dr. St. Louis, who is also with the Department of Biochemistry, University of Montreal. In many instances, the patient’s condition deteriorates so rapidly that clinicians can’t wait for a culture, or they don’t want to take chances so they prescribe antibiotics prophylactically, adding to the problem of antibiotic resistance.

In a presentation at this year’s American Association for Clinical Chemistry annual meeting, Dr. St. Louis stressed the need for better assays for diagnosing sepsis that have a faster turnaround time than culture. Unfortunately, while studies of numerous cytokines, markers of cellular activation, and acute-phase proteins have yielded useful information that provides pieces of the sepsis diagnosis puzzle, no single marker has emerged that can provide clinicians with all of the information they need to manage Systemic Inflammatory Response Syndrome and sepsis patients.

"The pathophysiology of sepsis is so complex that I think we’re going to have to use combinations of markers, but we’re still trying to determine what those combinations should be," Dr. St. Louis says.

Many physicians use the Acute Physiology and Chronic Health Evaluation, or APACHE, scoring system as a measure of illness severity in critically ill patients, but this is a difficult system to manage, Dr. St. Louis notes. "You must put in a combination of parameters that could be clinical, such as temperature and heart rate, or could include laboratory measures such as albumin and creatinine concentrations, and potentially even CRP levels." Each parameter is assigned a numerical value on a particular scale, and a score is produced. "The higher the score, the bigger the problem, the worse the outcome," he says.

In the future, Dr. St. Louis says, laboratories may have a similar type of system for sepsis markers. "The best approach is probably something like a score, where you combine markers that give you the necessary statistical information with very strong support. However, if we can get a good measurable parameter, we won’t need complicated scores," he says.

The laboratory component of sepsis diagnosis could also be managed in the future much like laboratory testing for chest pain patients is managed today, where several markers are considered at once and measured serially. A number of studies have examined the value of combining currently available markers like IL-6, IL-8, TNF-α, and CD64, and researchers are sorting out which markers work best at various time points on the sepsis cascade. "One study showed, for example, that combining IL-6 or CRP with CD64 provides a 100 percent negative predictive value that a patient has sepsis with 100 percent sensitivity," Dr. St. Louis says. Other studies have shown that a combination of CRP and IL-6 provides a better positive predictive value and negative predictive value for the diagnosis of sepsis than either marker alone.

But Dr. St. Louis cautions that studies of potential sepsis markers to date have demonstrated variability. "While it may seem that these tests are good either for diagnosis or prediction of outcome or severity, there are actually a lot of overlapping results, and you have a lot of variable correlation, so we really can’t use them to give us clear-cut answers. They are useful, but they’re not what we are really looking for," he says.

Thus, the search for definitive sepsis markers continues, and there’s no shortage of new assays to study. Last year, the endotoxin activity assay, a chemiluminescent test that measures concentrations of endotoxin, the toxic component of gram-negative bacteria, became the first test for identifying ICU patients at risk for developing severe sepsis to gain FDA market clearance (under de novo classification). Recently, findings from the prospective Multi-center Endotoxin Detection in Critical Illness (MEDIC) trial—a 10-center, multinational study that examined the relationship between endotoxin levels and patient outcomes—showed that endotoxemia is present in more than half of all patients admitted to the ICU (J Infect Dis. 2004;190:527-534).

These findings demonstrate that the endotoxin activity assay, or EAA, can identify patients at risk for developing sepsis within one hour of their admission to the ICU, says Paul Walker, MD, PhD, CEO of Spectral Diagnostics, manufacturer of the assay. "We found that endotoxemia is relatively common in patients in the intensive-care unit," he says. "The MEDIC study demonstrated that about 60 percent of all patients had elevated levels of endotoxin, yet less than 10 percent had an identifiable infection by culture." Endotoxin levels were elevated in the majority of patients who developed severe sepsis regardless of the type of organism cultured, he says, and therefore served as a more reliable biomarker for evaluating the risk of sepsis for all patients.

Significant levels of endotoxin in the blood may come from a focus of gram-negative infection, or by translocation from the large reservoir in the gastrointestinal tract. The latter occurs in patients "where the gut mucosa has been compromised due to a systemic insult such as shock," Dr. Walker says.

The study also demonstrated that when endotoxin levels were high, regardless of the source, there was a significant relationship between the levels of endotoxin and adverse outcomes associated with being in the ICU, including length of stay and mortality, he says. "The availability of this early marker of sepsis," Dr. Walker says, "can assist the clinician by identifying candidates for early goal-directed treatment of sepsis."

Soon, American researchers may have another new FDA-cleared sepsis assay to study. William Frank, general manager of Brahms Diagnostica, LLC, Norcross, Ga., says the company has submitted a 510(k) to the FDA for its procalcitonin, or PCT, assay, and expects market clearance in 2005. A relatively small biomolecule (12.6 kD), PCT is the hormonally inactive propeptide of the hormone calcitonin, which is normally produced in the thyroid. Because it is degraded by proteolysis in individuals with normal metabolism, PCT usually remains at undetectable levels (<0.1 ng/mL) in healthy people. But in severe infections caused by bacteria, fungi, or parasites, as well as in sepsis, PCT serum titers may increase to over 500 ng/mL.

"PCT is used extensively throughout Europe and is a well-recognized marker for bacterial infection and sepsis," Frank says. Brahms is still working with the FDA on an exact indication for use, but reference ranges have essentially been established. "If you do not have sepsis, your cutoff is probably going to be below 0.5 ng/mL," Frank says. If the PCT level is greater than 2.0 ng/mL, bacterial infection with systemic involvement is probable, and there is a high risk for progression to severe bacterial infection with systemic inflammation.

The scientific community generally accepts that PCT is useful in determining the severity and onset of sepsis, though the physiology behind why PCT is expressed by every cell in the body in sepsis is unknown. "There’s major literature out there on procalcitonin, and it’s clearly different in patients with sepsis compared to other patient populations," Dr. St. Louis says. PCT rises earlier in the sepsis response than CRP, and Brahms Diagnostica is positioning the assay as an early marker.

The company has two assays that are available now in the United States for research use. The first is a quantitative assay called PCT-LIA (luminescence immunoassay). It’s performed on a luminometer, and this is the format with which Brahms approached the FDA. However, company officials are speaking with U.S. manufacturers in the hopes of licensing PCT for other platforms, Frank says. A second PCT assay, called PCT-Q, will be able to provide rapid, semiquantitative results. "Within 30 minutes, you’ll have an indication whether the patient does or does not have a bacterial infection, and whether or not they should go to the ICU. In Europe, when you get to the point where you need to monitor the severity and the effect of antibiotic therapy, that’s when you would use the quantitative assay," Frank explains.

Those using PCT in neonates must be careful, however, because in the first 48 hours of life, PCT can be highly elevated, even in the absence of infection. "One theory is that the immune system in a neonate is developing at that point in life, and that might explain why PCT levels are high at birth and then fall over the course of several days," Frank says.

Unlike C-reactive protein, PCT does not appear to be elevated in the case of cardiogenic or viral infections, but Dr. St. Louis would like to see more studies. "We looked at PCT and CRP, performing a meta-analysis that whittled down relevant papers to only 12 really good, controlled studies with clear information value," Dr. St. Louis says. "What we found is that if you look at the summary receiver operator characteristic curve, that based on the data, procalcitonin appears to be better than CRP, but they both really seem to have variable utility as a diagnostic marker for sepsis. The threshold is somewhat variable."

Another new sepsis marker with an increasing amount of data behind it is lipopolysaccharide binding protein, or LBP. A recent study in Intensive Care Medicine (2004;30:1454-1460) showed that in critically ill neonates who are less than 48 hours old—the population in which PCT elevation is not necessarily related to infection—LBP is a better marker of sepsis than IL-6 or PCT and it is comparable to CRP. In critically ill neonates older than 48 hours, the researchers found LBP to be better than IL-6 and CRP, but comparable to PCT.

LBP binds to endotoxin or lipopolysaccharide from gram-negative bacteria to form a complex that can be recognized by the CD14 molecule, and is elevated early in an infection that could lead to sepsis in an at-risk patient. Gail Mastright, product manager for Diagnostic Products Corp., a company that manufactures an automated LBP assay, says, "When endotoxin is introduced into the body, LBP, an acute-phase protein, is produced by the liver as early as six hours after exposure." Ongoing studies conducted for the past few years in more than 200 ICUs by DPC Biermann, the company’s affiliate in Germany, have demonstrated that a combination of IL-6 and LBP assays can be used successfully to diagnose and monitor patients at risk for sepsis. In particular, Mastright says, "the German physicians report that they can distinguish local from systemic infections in ICU patients by measuring these two analytes daily." These assays are available now in the U.S. for research use, Mastright says.

DPC’s LBP is being evaluated now in the Genetic and Inflammatory Markers of Sepsis (GenIMS) study, a multi-center trial with 2,230 subjects enrolled. The Department of Critical Care Medicine at the University of Pittsburgh School of Medicine is coordinating the NIH-sponsored study, which seeks to identify whether certain alleles for key inflammatory molecules are associated with the risk of developing pneumonia and the risk of progression to severe sepsis, septic shock, organ dysfunction, and death. The researchers hope to find a pattern of inflammatory proteins in the blood that will lead to the development of a new sepsis marker. Though DPC has been given only preliminary data from the study, Mastright reports that LBP shows potential in detecting when a septic patient is becoming severely septic.

Also on the horizon is an assay Beckman Coulter is developing, called inducible nitric oxide synthase, or iNOS. "In sepsis, inducible nitric oxide synthase is produced when there are co-factors present in circulation which include cytokines and LPS or portions or remnants of matter from organisms like a bacteria or fungus or virus," says Linda Felgen, director of market development for the Immunodiagnostic Business Center at Beckman Coulter.

Though there’s still much to learn about iNOS, Felgen says, it appears that it’s produced in sepsis in tissues like the neutrophils and the endothelium. "Until recently, iNOS was believed to be entirely contained within cells and not free or soluble in circulation," she says.

iNOS induces production of nitric oxide, which plays an important role in the sepsis cascade, Felgen says. "In sepsis, over-production of nitric oxide drives the inflammatory response out of control, resulting in hypotension, altered renal and liver function, and coagulation abnormalities, potentially leading to death through multiple organ failure," she says.

Based on the company’s data, iNOS appears to be a very early marker that may rise many hours before an elevation of white cell count and temperature. Felgen says Beckman Coulter researchers are learning more every day about the molecule. The company aims to move the assay into clinical evaluation in 2005.

Beyond 2006, clinical laboratories may see assays like adenosine deaminase binding protein-26 (ABP-26) arriving on the scene. Released by the nephrils into the urine when a patient is suffering from acute renal failure, ABP is a protein released by brush border epithelial cells that has been shown to rise in urine four to seven days before creatinine rises in serum. "The release of the adenosine binding protein is a direct result of the damage to the nephrils usually caused by infectious insult, whereas endotoxin is clearly a release of endotoxin from the damage of a cell," says Richard Gill, CEO of Signet Laboratories, the company that makes the ABP-26 research-use-only assay.

ABP-26 is tied to two therapeutics that Signet will also be developing, and Gill expects that the test will be needed to titrate the therapies. "The RUO that we launched was with a urine standard, and we’re in the process now of developing the assay with a recombinant protein standard, and that should facilitate the use of it," Gill says. The company is looking for investors to take the drugs and the assay into trials.

While labs wait for the ultimate sepsis assay to arrive, whatever it may be, research into multiple markers continues, and those interviewed for this article say sepsis is a condition that is likely to require a suite of markers. Says Felgen, "If you look at gene expression arrays over the course of a patient’s condition starting from little or no symptomatology and ending in severe sepsis and septic shock, what you see is that the patterns on those arrays change very quickly and dramatically over the course of the condition. This tells us that genes are upregulating and downregulating at different times in different combinations, so we should assume a variety of biomarkers will be necessary to characterize the patient’s condition." In her view, a group of biomarkers will be needed to manage this disease, some for early detection, some to confirm the source of infection, and others to assist with therapy choices.

For now, Dr. St. Louis says, clinicians are not comfortable with the markers they have for sepsis, either for diagnosis or prognosis. "We have tests we think are useful, some of the new approaches we are taking may be good, and many of the biomarkers are easy to measure, but we just aren’t there yet," he says. "If we can define the particular patient population we are targeting more precisely, we might then be able to better identify the utility of a particular test."

Better-designed studies are also key, and Diagnostic Products Corp. recently standardized its cytokine assays to World Health Organization standards so that researchers can be assured that studies using their assays will be comparable. The bottom line, however, is that clinicians need markers for sepsis that will provide objective numerical data, Dr. St. Louis says.

The Society of Critical Care Medicine projects that the incidence of severe sepsis will rise from 750,000 cases in 2004 to nearly 1 million cases in 2010, making the need for good sepsis assays even more critical. Says Felgen, "I can’t think of a clinical diagnostic dilemma that needs more attention from our industry than sepsis."

Sue Parham is a writer in Edgewater, Md.