College of American Pathologists
Printable Version

  Feature Story


cap today

Latest assay opens another sepsis frontier

May 2003
Karen Titus

Diagnosing sepsis has become an open-and-shut case. Except for the “shut” part.

Sepsis opens and shuts with all the steadfastness of saloon doors in a Hollywood western that flap wildly with every card-playing drunk who’s tossed through the portal. “Is the field becoming more clear? Yes, it is. It clearly is,” says John Marshall, MD, FRCSC, FACS, professor of surgery and director of research for the Interdepartmental Division of Critical Care at the University of Toronto.

The topic being sepsis, there’s a catch. “What’s becoming clear is just how incredibly complicated the whole field is,” says Dr. Marshall, who is also head of the Division of Clinical Investigation and Human Physiology at Toronto General Hospital.

Ten or 15 years ago, Dr. Marshall says, physicians and researchers assumed sepsis diagnosis and treatment would turn on one or two key molecules. Since that has utterly not been the case, the medical community has been forced to take on sepsis from innumerable angles, ranging from coagulation to relative adrenal insufficiency, from low-stretch ventilatory strategies to enhanced insulin therapy.

Then there’s endotoxin. The Food and Drug Administration recently allowed Spectral Diagnostics’ Endotoxin Activity Assay, or EAA, for identifying ICU patients at risk for developing severe sepsis, to qualify for market clearance under de novo classification. With this, the sepsis door closes and opens yet again.

Endotoxin is not some breakthrough substance; it’s been around
long enough to have its own mavens. Steve Opal, MD, professor of medicine at Brown Medical School, Providence, RI, calls himself one of them. “I’ve spent all my life in the laboratory studying endotoxin, so of course I’m very excited about this assay. But I’m a bit biased as well,” he says. “We’ve been looking for many, many years to find a rapid diagnostic assay that can assist us in determining whether a patient’s going to be high-risk or lesser risk for developing complications.”

Researchers have long recognized the pivotal role of endotoxin in sepsis but until now have been unable to devise a reliable, real-time method of measuring it. The EAA, which measures circulating endotoxin in the blood, therefore elicits no small amount of excitement among its developers and early users. “We’ve never been able to do that before,” says Dr. Marshall, who served as the lead PI for the recent MEDIC trial, which evaluated the EAA. Results of the trial are under review for publication.

The Multi-center Endotoxin Detection in Critical Illness trial was a multinational study, with 10 investigative sites in Canada, the United States, the United Kingdom, and Belgium, reports Debra Foster, BSC, director of clinical studies at Spectral Diagnostics and coordinator of the trial. More than 1,000 patients were enrolled in this observational study. “We looked at outcomes for these patients in terms of levels of endotoxin and that association with outcomes, such as ICU length of stay, ICU mortality, association with APACHE II score, and organ dysfunction—and, of course, severe sepsis, which was our main study objective,” Foster says.

The assay has traveled a long developmental road, says Paul Walker, MD, PhD, president and CEO of Spectral Diagnostics and a former critical care specialist.

The endotoxin molecule is present and very well preserved in all species of gram-negative organisms. Common to all endotoxin molecules from different bacterial species is lipid A. “The lipid A moiety is common, but it has a single epitope,” says Dr. Walker—which effectively rules out routine ELISA-type technology for its analysis.

The EAA is a chemiluminescent method that uses the CR1 and CR3 receptor-induced oxidant production of polymorphonuclear leukocytes as its detection platform. Specificity is achieved by the differential priming of neutrophils in whole blood by lipopolysaccharide-antibody complexes. The control measures baseline whole blood neutrophil oxidant production, and the test assay contains murine monoclonal IgM antibody against the lipid A epitope of LPS and measures the enhanced chemiluminescent response of the neutrophils in the presence of LPS-antibody complexes. (These and additional assay specifications were reported in the Journal of Immunological Methods [Romaschin AD, et al. 1998;212:169-185].)

“What we’ve been able to do is figure out a homogeneous assay that would make use of its characteristics and allow us to accurately measure the lipid A activity,” Dr. Walker says. The tail length and therefore weight do not relate to the severity of endotoxin’s response “So that’s where there’s been quite confusing results related to endotoxin measurements in the past.”

The assay’s use of whole blood is a limitation and a benefit, Foster suggests. Though that limits the time to run samples to a three-hour window, “it’s really meant to be a rapid test,” she says.

Results are available as quickly as, if not quicker than, a Gram stain, says Dr. Marshall. “It certainly is substantially quicker than blood cultures, because you can have the results back within 60 minutes, rather than 48 to 72 hours.”

But acting on that information flings the door open again. Wide open. Since physicians have not routinely been able to tap rapid endotoxin results, “it’s hard to say we should do such-and-such with them. We just don’t know yet,” says Dr. Marshall. “We need more data on whether or not it [the assay] can be used to diagnose and track infections. Obviously the limiting factor there is the fact that endotoxemia is as common as it is.”

As one MEDIC investigator points out, conditions other than pure gram-negative infections will cause increases in endotoxin and result in a positive assay. “I think when we initially did the study, we were hoping that it would be very sensitive and indicate that patients had a gram-negative infection. But it really doesn’t perform like that,” says Edward Abraham, MD, vice chair, administrative affairs, for the department of medicine and head of the Division of Pulmonary Sciences and Critical Care Medicine, University of Colorado Health Sciences Center, Denver.

Instead, endotoxin may function as one of several warning lights for physicians, indicating a patient is at greater risk of an adverse outcome, though, like Tom Ridge’s color-coded warning system, the alert doesn’t necessarily spell out a clear course of action.

One hypothesis that could be tested over the next few years, Dr. Marshall says, is the notion that the measured endotoxin derives from the patient’s gastrointestinal tract, which might give the assay standing to monitor gut function in critically ill patients.

The assay would also come into play as specific antiendotoxin therapies are developed. According to Dr. Marshall, “There are at least four different strategies that I’m aware of that are in the early stages of clinical development.” Since there’s no shortage of methods for neutralizing endotoxin, “it’s certainly something that will evolve,” he says.

Dr. Opal also predicts the assay will come into its own eventually. It will be especially useful for clinical studies, he says, as therapeutic agents are developed. These could include E5564, a lipid A antagonist now in clinical trials, and the anti-CD14 monoclonal antibody.

He, too, sees prognostic possibilities for the assay, including value as a negative predictive study. “It might be able to help exclude patients who have gram-negative infections, and might be able to identify people who are at greater risk to get gram-negative infections,” he says. “So if the endotoxin test is negative, the chances you have a concomitant, ongoing gram-negative infection are very low. That’s useful information.”

In the clinical trials for the assay, moreover, patients who had endotoxin in the their blood were more likely to be sick and be in the ICU longer than patients without circulating endotoxin. “That will be of interest to see if that can be confirmed in a larger number of patients in a variety of different situations,” Dr. Opal says.

The assay may direct more effective use of current treatments as well, says Dr. Walker, pointing to a New England Journal of Medicine article (Rivers E, et al. 2001;345:1368-1377), which suggests that early goal-directed therapy—adjustments of cardiac preload, afterload, and contractility to balance oxygen delivery with oxygen demand—significantly improves outcome in ICU patients with severe sepsis and septic shock. “Knowing that a patient is clearly at high risk for developing sepsis will alert physicians to begin their treatment more quickly, and will also be useful in identifying patients who are most likely to benefit from any sepsis treatments, either specific antiendotoxin therapies or other modalities,” Dr. Walker says.

But he sounds a final note of caution when he says, “Sepsis is littered with the corpses of companies that have come forward with a sepsis therapy, without laboratory tools for identifying patients more likely to benefit from treatment.”

Other questions continue to pop up in sepsis, the chief one being, as Dr. Opal puts it, “What’s going on?”

“We know a lot about how endotoxin causes disease processes in animals, and we know a lot about how endotoxin at low doses causes reactions in human volunteers,” he says. What’s not known is the pathophysiology that’s directly attributable to endotoxin release in septic patients. “There’s still an open discussion, and a very active one, about the relative role of endotoxin-mediated pathologies in combination with other events that are not endotoxin-mediated in determining the outcome of the patient,” Dr. Opal says. Answers may emerge with the development of very specific, highly potent antiendotoxin drugs, which would allow researchers to remove endotoxin from the mix.

Endotoxin is sufficient to cause sepsis, but not necessarily required to do so—patients may have fungal infections or gram-positive infections, for example. “So it’s certainly conceivable that patients may develop severe sepsis and even die from septic shock without having endotoxin in their circulation,” Dr. Opal says.

“We know that if you give endotoxin to normal persons, they develop a very impressive illness that looks for all the world like sepsis,” he continues. “And we know that patients who have very high levels of endotoxin in their circulation tend to do very poorly in sepsis. But there’s always been this question of, 'is this a marker of disease severity, or is it a cause and effect?'” Does the permeability of the intestine increase as patients become sicker, causing endotoxin to weep from the lumina of the gastrointestinal tract? Or is the endotoxin itself causing the severe illness?

One need look no further than Spectral’s Dr. Walker to comprehend just how outlandishly complicated sepsis is. Though he has many good reasons to tout his company’s new assay, he remains as guarded as anyone in trying to pinpoint its value. “This will be part of these patients’ assessment,” he says. “It won’t be all of their assessment. It’s not as if you’ll have one test, and everything else will be straightforward.”

Dr. Walker terms endotoxin a proximal mediator of sepsis. “It’s likely that it’s one of the first mediators that’s liberated in the sepsis cascade,” he says.

It’s barely necessary to mention that sepsis is shot through with other critical elements. Chief among these is the hypercoagulable component, which may be the target of activated protein C (Xigris), a relatively recently approved drug that has been shown to significantly reduce mortality in patients with severe sepsis.

But this has its own set of nagging questions, few of which are being answered quickly.

“There have been two studies, basically, that have come out since the Xigris study was completed, one with antithrombin III and one with tissue factor pathway inhibitor,” says Dr. Abraham. “Both of those studies were negative. So I think for most of us, it implies that the anticoagulant aspects of Xigris are not why it works. It may actually work primarily because of other characteristics: anti-inflammatory, for example, or its effects on endothelial cells.” Dr. Abraham is the principal investigator of the ADDRESS study, launched by Xigris’ maker, Eli Lilly. The study will investigate the drug’s effectiveness in 11,000 severely septic adult patients who have a lower risk of death, across approximately 1,000 clinical trial sites worldwide. And while Dr. Abraham’s interest in both Xigris and endotoxin is robust, he, like many other sepsis experts, is also looking ahead toward genomics-based approaches to sepsis. “There’s no question that these types of analyses will be coming along soon,” he says.

“Genomics- or PCR-based rapid diagnostics for bacteria in the bloodstream could be extremely useful in the near future,” agrees Dr. Opal, emphasizing that clinicians desperately need help from laboratories in quickly identifying patients who are septic or bacteremic. “We’re very frustrated to have 30 to 40 percent of patients who enter these clinical trials in whom we can’t make a specific microbiologic diagnosis,” he says.

There’s also no question of the need for more specific markers, since drugs such as Xigris are being delivered to a broad, heterogeneous patient population with infections from multiple microorganisms. “The real frontier in the future will be with diagnostics that will tell us who is very likely to respond to these drugs,” says Dr. Abraham.

Sepsis is surrounded by frontiers. One of the most recent innovations
in clinical sepsis research has been the use of corticosteroids in severe sepsis patients. A multicenter study done in France and published in JAMA (Annane D, et al. 2002;288:862-871) showed that patients with vasopressor-dependent septic shock appear to benefit from a one-week course of stress-dose steroids if they have relative adrenal insufficiency, as determined by an ACTH test.

The effect of the study, says Dr. Opal, has been to refocus the steroid debate, which has generated huge amounts of controversy for decades. In his view, the Annane study shows, fairly convincingly, that giving low-dose steroids—not megadose steroids used in the past—appears to benefit at least those patients who seem to have an inadequate adrenal reserve in response to septic shock. What’s not clear is whether the value comes from administering glucocorticoids, mineralocorticoids, or both—the trial added a low dose of the latter without a separate arm looking at glucocorticoids alone.

The ACTH test has its own problems, turnaround time being one of them. “Our current practice in patients in whom we suspect this problem is to administer steroids after having done a short-course ACTH stimulation test, but it’s 24 hours before we get the results of that test back,” says Dr. Marshall.

Dr. Abraham points to another dilemma. “There’s a question of what really defines adrenal insufficiency,” he says. A recent New England Journal of Medicine article (Cooper MS, et al. 2003;348:727-734) that addressed adrenal abnormalities in critically ill patients suggests that patients who had serum cortisols greater than 34 µg/dL were unlikely to have adrenal insufficiency. The Annane article, on the other hand, maintains that the patient’s baseline serum cortisol is not the critical issue; what counts is the amount it increases after treatment. Says Dr. Abraham, “There’s still a lot of controversy in terms of defining who has relative adrenal insufficiency among these septic patients.”

Less controversial, perhaps, is yet another recent approach: using enhanced insulin therapy to maintain euglycemia in septic patients.

There’s nothing new about keeping ICU patients’ blood sugar under control but, as Dr. Opal suggests, caregivers may occasionally give it short shrift. “These patients are very ill, and there are so many things going on with them. They’re getting hyperalimentation, and there’s many reasons for them to have an elevated blood glucose. In the past, we usually have not worried when a person’s blood sugar was 150 or 200 mg/dL.”

Researchers have recently shown, however, that stringently controlling blood sugars to keep them within euglycemic ranges results in decreased instance of infection, shorter ICU stays, and improved outcomes. “It’s not doing anything that we didn’t already do before,” says Dr. Opal. “It’s not anything new, innovative, or high tech. It’s simply using what we’ve been taught and already know, showing that if you do it correctly and rapidly and accurately, with specific protocol goals in mind, that patients benefit.”

There is little doubt that this mix of the old and new offer the best hope of taming sepsis. By relying on current crude clinical measures alone to make a complex diagnosis, Dr. Opal says, physicians confer the sepsis label on patients who have remarkably dissimilar pathophysiologies. “We may do them a disservice in a way, because we can’t determine the optimal treatment strategy.”

The lack of good, discrete diagnostic tests for sepsis has plagued the field for too long, he says, which is why he welcomes the endotoxin test. “I think it’s going to be a useful device.”

But, he quickly adds, “I don’t wish to give the impression that I think this is an open-and-shut case.”

Not to worry—no one, but no one, thinks it is.

Karen Titus is CAP TODAY contributing editor and co-managing editor.